DAW Build "Gotchas"
When I first decided to go for a computer that I could use for serious multi-track recording, it was a pretty big deal. At that time, most off-the-shelf home systems could not handle that job, so I ended up going to a professional to have a proper machine built. This was a good move, and I still recommend going to someone who specializes in DAW building if you have more money than time. Over time, though, the capability of relatively “ordinary” hardware has gotten much better, to the point where many of the readily available “package” systems have the processor power and disk speed to handle at least 24 tracks at a time. It is now much easier to put together a decent DAW machine than it was several years ago, and I built up my last two DAW systems myself (note: while I am a fairly competent computer user, I am not by any stretch an IT professional).
Invitation to Adventure
These days I feel fairly comfortable recommending that anyone who can be trusted with a screwdriver, is willing to read directions carefully, and take their time choosing and assembling components, consider building their own system. Of course, you still have to take the standard precautions necessary for working with any computer hardware, such as avoiding static discharge to the parts, so IMO the terminally careless need not apply. That said, I would like to offer a little guidance in system building so that you can avoid the “gotchas” that wait to trap the unwary. Like everything else high-tech, computer technology is a moving target, so for the most part I will not recommend specific components, because my recommendations would probably be out-dated in 6 months or less, so in this article I am mostly sticking to general principles. If you want an example with specifics, you can look at the article “My 06 DAW”.
In an ideal world, you would be able to walk into a store, tell them what you are doing, and have them recommend what you need. Unfortunately, a DAW machine is “specialized” enough that most computer stores will not have anyone who understands your needs well enough to advise you. If you go into a music retailer, they might offer packaged systems for music work, but these systems are often priced at a premium that is high enough that, for the same money, you could afford to go to a reputable system builder and get something customized to your personal needs. For the most part, computer store guys don’t know music well enough, and music store guys don’t understand computer hardware well enough.
To make matters worse, the music store guys often do not know the difference between the better products and those that have weaknesses that you may want to avoid. This is especially true when it comes to computer audio interfaces. It just so happens that your choice of audio interface affects what kind of computer you want to build, and the quality of that interface has a real effect on the quality of work that you will be able to do.
Naturally, the first aspect of “quality” that we think of here is sound quality, which is determined mainly by the sound of the converters you use. Many sound interfaces have built-in converters. This is true, for example, with the built-in sound hardware that comes with most computers, or the Sound Blaster stuff, or almost everything that is currently aimed at the “consumer” market. Some of the sound devices you can buy have very good converters in them, while the sound quality of others may not be all that great. Many multi-channel sound interfaces, especially those able to handle more than 8 channels, are purely digital interfaces, which must be connected to external converters. Some devices have internal converters and the ability to work with external converters as well (usually as a way to get more channels). Using external converters allows you to decide for yourself the type and quality of converters that you will use, and will allow you to upgrade later as your budget allows without having to replace the sound card/interface in your computer.
The other aspect of “quality” is how fast and how reliable the audio interface is. When we say “fast”, we are talking about a sneaky thing called “latency”. Every digital audio device made takes a certain amount of time to do its job, which causes a delay in the signal, and the amount of that delay is the latency of that device. If the latency is small enough, you can usually ignore it, but too much latency can sometimes make a job difficult or, in some cases, imposssible. As for reliability, what we are after if for the device to first to work properly, and second to work exactly the same way every time.
It turns out that one of the things affecting reliability here is the choice of interface type. Internal PCI cards have been around long enough for the bugs in that particular standard to be worked out. USB seems to be workable for the smaller interfaces with 2 or only a few channels. The recently popular Firewire interface, however, sometimes runs into problems. With some setups, there may be occasional glitches in the audio or other failures. Not all of the bugs in the implementation of this standard have been worked out, apparently. For that reason, if you are looking at a Firewire setup, try to find out how it is working for other people, and particularly using the computer system you have in mind. Be aware that some of the Firewire port hardware in computers does not work as well as on other setups (so far the word is that not all of the chipsets made to implement Firewire work equally well). If you get such an interface and plan to do serious work for paying clients, do careful testing before using it for a critical session. The physical build quality of the interface also affects reliability, but there is yet another element of the system involved...
The biggest hidden “gotcha” with audio interfaces is the quality of their drivers. A “driver” is a special program that tells the computer how to communicate with a piece of hardware. Your computer has drivers installed for all sorts of things, like the video card, your keyboard, your mouse, your network interface (NIC), the controllers for your hard drives... and the list goes on. Every sound card or device has a driver of its own that must be running so that the computer can work with it. A well written driver is essential to the quality of work that you can do with your sound hardware. A hardware device that should have real promise can be made completely unusable by a poorly written driver. In severe cases, bad drivers can cause system crashes, at worst resulting in the Blue Screen Of Death. More commonly, poor drivers are the cause of excessive latency, imposing inconvenient system delays that are especially important if you want to monitor input sources through your DAW.
Sadly, nothing on the box the sound device comes in will tell you whether its drivers are really good. Some of the more reputable manufacturers a pretty consistent in providing good drivers for their products, but to be really sure you probably want to check with some other users to see what they have found that works well for them. A good place to find this information is in the various user groups on line, especially those for the DAW software that you use or for the sound hardware you want to buy.
The part of the job that requires the most care is the selection of components. This starts with deciding exactly what you want out of your new machine. There are some possible tradeoffs involved here. For example, I personally set a priority on ease of portability, so I chose to build a machine with minimal hardware options, which allowed me to keep the case small. If you want, instead, the have multiple hard drives in the machine and expect to add multiple accessory cards (such as hardware DSP cards, which these days are a pretty popular option), you will likely want a larger case. Make sure you have some idea what you will want for about the next 3 years or so (until you decide it is time to built your next “newer and better” machine), and allow for future expansion if you think you might need more later than what you can afford now. You probably cannot build a machine that is the best in every possible way, so make sure that you build the machine that is best for what you need.
The first part you must choose is probably the processor. Almost any processor now offered in a new computer has enough power to do worthwhile audio production, especially if your track count doesn’t go too high. Until recently, though, my advice was always to choose a processor that was at what I call the “knee” of the price/performance curve. Generally, as performance goes up, so does price, but there is a point beyond which a fairly small increase in performance demands a large increase in price; I would usually buy just below that point. This still makes sense if you are anxious to get as much power in your system as possible, and always a good idea if your budget permits, but today you can get a pretty decent amount of processing power at a surprisingly modest price. For my last DAW, I spent only about $130 on the processor, which was somewhat faster than the processor for my old machine, and that earlier processor had cost me almost twice as much.
There is some argument about whether Intel or AMD makes the best processor for DAW use ( I have used both with good results), but the fact is that you can get very good processors from both companies, and the competition between the two has certainly been a major factor in getting us all more value for our money.
The one feature that I always look for in my choice of processors is on-chip memory cache, which helps the system run faster by reducing the amount of access needed to the main memory on the motherboard, thus enhancing system speed. Not all processors have it, but I make sure to get the largest cache as I can get in a reasonably priced processor.
After you have chosen a processor, the most central decisions you make will be about the motherboard and the case. If you are not worried about the size of the computer you are building, choose the motherboard first and then find a case that fits it properly. This means that you can select a motherboard with any features you want. Certain combinations of built-in features simply are not available in smaller motherboards. The most obvious factor dictating motherboard size is how many card slots you need for add-ons. If you plan to use an internal sound card, you will need to have a slot available for it. If you are going to add a special video card, you will need a slot for that, too. Any special feature, such as a network interface, a firewire port, a special drive controller, or a hardware DSP card will need a slot also.
If size is a consideration, the type of case you choose may limit your available motherboard choices. Today there are several different standard “form factors” for motherboards, each with certain dimensions and mounting requirements. The case manufacturer will tell you what form factor motherboard a given case is designed for, so if you absolutely must have a certain specific model and style of case you will have to use a motherboard that fits that form factor. For example, for my last two computers I wanted something that I could manage to carry under one arm. In both, er, cases, I chose mini-tower cases made for motherboards with the “micro ATX” form factor.
Another factor in motherboard choice is what processor you will use. You need to know what socket or slot type your chosen processor uses, and you must buy a motherboard that uses that particular socket. It is probably best not to build a system with a socket type that is about to become obsolete, since it is likely that you will want to upgrade the processor at least once before you bulldoze the system (I have done this with at least 3 DAW machines in the past).
Next is your choice of on-board features. To begin with, make sure that the motherboard can handle as much memory as you will ever want for the system (these days I would consider 2 gigabytes of memory a minimum, and this number will only go up over time). You also probably want to look for a system that uses faster memory (again, these numbers keep going up, so I won’t recommend a specific speed). Now we get into the various options. One of these is on-board video. The good news here is that a DAW does not normally need really cutting-edge video features. If you are only using a single video display and not doing any “extended desktop” tricks, almost any standard on-board video setup should work just fine. It may be a good idea, though, to have a video card slot available in case you may want something better in the future (I have taken that step in two different machines so far, where I started out using onboard video and upgraded to a dual-monitor video card later). I strongly recommend you get onboard LAN capability as well. In fact, I usually try for as many on-board features (of those I will use) as possible for one specific reason: with all the hardware and drivers for it coming from the same manufacturer, conflicts between various devices and drivers are much less likely, which may well mean less troubleshooting later.
Another option to consider is USB2 or Firewire capability. It might be nice to get both, but if you have to choose one or the other, USB2 is more widely useful since there are so many USB devices out there. At present, for Firewire capability I would probably choose an add-on card. It turns out that some of the available Firewire chipsets work better than others, and using an expansion card for this purpose allows you to replace it if you got the wrong type the first time around. Careful choice here will be especially important if your audio interface uses a Firwire port.
Onboard audio is usually not particularly important, although it may not be possible to get the other “integrated” features you want without getting onboard audio as well. My last 3 DAW machines came with it, and I usually don’t use it at all. There is, however, one potential good reason for using on-board audio: system sounds. If you like having system sounds available for certain types of alerts, it is wise to send those sounds to the onboard audio and feed them to their own small speakers. That way, if you happen to be monitoring a mix at high volume, you won’t get blasted with a surprise “DING!” or other obnoxious Unexpected Loud Noise.
One special caution about motherboard choice: In the past, not all motherboards would work with pro audio sound cards. I learned this the hard way when trying to make an old Sonorus StudI/O card work with computers from both HP and Dell, and in both cases there was no way to get the card working (although the same card DID work in another machine). In at least one case, the problem was that the motherboard did not allow for reserving and IRQ for a specific PCI slot. As a result of that experience, I do not ever assume that any untried motherboard type will work with the sound card I use. Fortunately, this was not an issue with the last 2 machines I built.
There are a few aspects of your choice of case that you may want to think about. The first, of course, is to be sure that it is compatible with the mounting and space requirements of the motherboard, not to mention space for any accessories you may add. Next is the power supply. One thing to watch out for is whether the power supply is fully compatible with the motherboard and processor you have chosen. New standards keep coming out, and not all of the supplies currently sold meet all of the latest standards.
Then of course there is the capacity of the supply. If the supply is too small, you may be in for trouble. In fact, it can be a good idea to have a supply a bit more powerful than the minimum that is recommended. It is nice to have power in reserve for future expansion, but there is also another reason: added heat dissipation capability. The designers of a power supply know its efficiency, and thus how many watts of heat it must be able to “throw off” under full load. Under any smaller load, there is less waste heat, which puts less of a burden on the cooling capacity... so the supply may actually run cooler.
Keeping Cool and Quiet
One of the changes I made to my last computer after I built it was to replace the power supply with one that was designed to be much quieter. The noise, of course, comes from the fan. On most power supplies the fan is right at the vent to the outside, so any noise it makes projects directly out to the open air. On the replacement supply I used, the fan is mounted facing the inside of the computer, thus using the contents of the supply chassis as well as the computer case to help muffle the noise it makes. It is also a larger diameter fan which thus can move the same amount of air at lower RPM than most supply fans. The biggest noise reducing factor, though, is that instead of running full speed all the time, the fan can be slowed down. The supply has a switch that allows full speed, temperature controlled speed, or a special “low noise” mode that keeps fan speed as low as possible. In that last mode, the supply is very quiet. Here, then, is where the benefit of an “oversized” supply really comes into play: I put in a 400 watt supply, but the way I built the computer it seldom pulls more than about 150 watts (I used a wattmeter to measure the power draw). This means that the heat dissipation capability of the power supply is far more than what the computer will demand in normal use. With less power dissipated as heat by the power supply, less airflow is needed to get rid of the waste heat, so the fan does not need to run at high speed, which means that the supply can stay quiet.
Next there is the consideration of adequate air flow so that the processor and other components on the motherboard can stay cool enough for safe operation. Intel considers this factor important enough that, last time I checked, they actually listed certain cases that had been tested with their motherboards and found to have sufficient air flow for adequate cooling. There are actually two parts to this requirement: one is that there must be the ability to move enough air, and the other is that the air motion must reach the right places (meaning where the most heat is generated). In many computer cases, the internal airflow design is such that internal components are actually kept cooler when all covers are in place than when the covers are removed. One way to check for yourself whether your computer is adequately cooled is to use software that reads the temperature sensors that are built into the machine (most motherboards have these sensors now) so that you can tell how hot things are getting. Some motherboard manufacturers, including Intel, supply programs that will display temp readings for you, and there are also a few freeware monitoring utilities available on line that allow you to make such tests.
Another strategy for keeping things cool is to have fewer things that generate heat included in the computer, and to use, as much as possible, parts that generate less heat. For example, I tried to choose a processor that would generate less heat and so not need as much air flow. The memory modules I bought included metal heat sinks to help throw off heat more effectively.. The video card I bought had just enough power and features to do what I needed, and not much more.
As a general rule, more powerful or faster hardware tends to generate more heat, which means it has greater cooling requirements. Today’s fastest and most powerful video cards almost all have their own cooling fans to deal with the heat they generate, and those fans make more noise. The more exotic video features that you can get now are used more for video games than anything else, and DAW software generally does not need anything all that fancy. For my use, then, I reasoned that the cheapest dual monitor card that I could find would probably do all I needed, which turned out to be true. Being a “less powerful and fancy” model, there was less heat to worry about, and this particular card used a big metal heat sink instead of a fan, which also means that the card is silent.
While we are talking about video, I should mention that not all video cards are well suited to DAW use. The drivers of some video cards can cause conflicts with audio card drivers, resulting in unpredictable glitches in the audio. In some cases you can fix the problem by reducing the hardware acceleration. This setting is often found under “advanced” properties for your video settings in Windows. Some video hardware, though, simply will not play nice with your audio hardware and drivers. If you are not sure about your choice of video card, you should check with the manufacturer of your chosen sound interface, and ask around in one of the user groups, either for your sound hardware or for your chosen DAW software. I personally have had good experience with ATI video hardware (there is other stuff that will work, I just haven’t tried it).
Another source of heat in a computer is the drives it uses for storage. At the very least you need one hard drive. In fact, for an audio machine you can often get along just fine with a single internal HD. In my newest machine, that is the only drive I installed. I decided that I did not need a floppy drive at all. I also chose not to install any optical drive. Even for installing the operating system, I used an external USB CD-ROM drive. Not including anything but the single hard drive meant fewer heat sources, thus less heat to get rid of.
By being really “stingy” about what I put into my computer, I managed to keep the heat dissipation low enough that I do not need to run an added case fan, and I can keep the power supply fan at the absolute minimum speed. I still have the processor fan, of course, but the fan that came with my “retail box” AMD Athlon64 3200+ processor, although it is the loudest thing in the box, is surprisingly quiet, leaving me with the quietest DAW I have ever had.
Once you have chosen your processor, motherboard, and case, the worst of the battle is really over. Everything else that you buy has to fit your motherboard and your case, which helps limit the number of choices you have to make.
Next most important is probably the choice of hard drive, although these days it is difficult even to find a drive that is too small for a DAW, and the performance of most new drives is more than enough to keep up with reading or writing the files for 24 or more tracks at once. Just about any 7200 RPM drive is fast enough. You could go for a 10,000 RPM drive, but the additional heat and noise of the faster drive is not wanted, and the added performance is probably not needed. You want to have a decent cache on the drive as well. I prefer to get drives with at least 8 megabytes of cache, which can make it easier for the drive to “read ahead” enough to keep up with the demands of many tracks/files being accessed at the same time. One drive should be enough, unless perhaps you are running video and audio at the same time, and even then you can probably use an external USB2 or Firewire drive instead of further burdening the power supply and cooling fans of your computer.
The Soft Options, Beginning With OS
Once you have your hardware together, it’s time for software installation. If you want to get the best possible performance from your DAW, it is important not to burden the machine with stuff you don’t need for audio production. This starts with how your operating system is set up. I personally still use Windows 2000, although I know that Windows XP can also be configured properly for DAW use. (As of this writing, I have reason to suspect that Windows Vista does not perform as well, and there is definite information that Vista requires more powerful hardware to get the same level of performance as Win 2000 or XP.)
Lately there is more interest in the use of Linux for audio production, and there are some applications now under development for it that show some promise, but the refinement of features and power in this area is going to take a while to catch up to the offerings available on the Windows platform. Some Windows applications can be run on Linux by using an add-on utility called WINE, but this too is still something of an experiment. You also need to make sure that there are proper Linux drivers available for your sound hardware. In some cases, the hardware manufacturers do have Linux drivers available, but sometimes those drivers do not fully implement the features of the hardware, and some software that comes with the hardware has not been ported to Linux at all. I have never done much experimentation in this area, and so cannot offer detailed advice about setting up a Linux-based DAW.
OS installation starts with partitioning your hard drive. I would suggest setting up a minimum of 2 partitions: one for operating system and program files, and another for all of your sound and other project files. Given the large amount of space on drives these days, it makes sense to make your OS partition 10 Gb or larger. Format the boot partition using the default settings, but for the “big file” partition, set the allocation unit (sometimes called “cluster”) size to 64k. This is more efficient for large files (and most audio and other project files are, by comparison to text files and such, quite large), reducing the amount of disk seek time when you are using these files and probably cutting down somewhat on file fragmentation.
The rest of the OS installation process is pretty much the standard stuff. After installation, though, you may want to make some adjustments. In particular, you will want to eliminate the running of programs and services that you don’t need.
If getting maximum performance from you DAW is important to you, you want to dedicate the machine for ONLY that use. This leads us into the important issue of
It is especially important that you do not use your DAW for any on-line stuff such as e-mail or browsing the Web. In fact, if your DAW is going to be on a network the network itself needs to be made secure from outside interference. This has to be worked “from the outside in”. Most likely, if you are following my recommendations, you will use at least two computers, one for the DAW and another for just about everything else. In my studio, I have a DAW machine and an “office” machine, and these are both on my home computer network The computers on the network all connect to a common router/switch which acts as a firewall. Any online traffic between my cable modem and any computer must pass through this firewall. Because of the way my firewall is configured, any outside attempt to probe my computers finds NOTHING. No outside intrusion can reach the computers on my network without “cooperation” from the inside.
While a firewall is essential for proper network security, it is, unfortunately, not enough. My wife and son have both had virus attacks on their computers, and have had to install virus scanners and other security software for protection. My online machine also gets scanned periodically for viruses. Ideally, any machine that goes on line will have a full set of protection software installed and running. Naturally, this means that a certain percentage of overall processor power is taken up with security measures. Unfortunately, on a DAW machine this can mean trouble.
In order to insure that the performance of your DAW is not limited by software conflicts, you must NOT install any security software on your DAW. This also means that in order to keep your DAW secure, it must NEVER go on line. Any transfer of files to or from the on-line world must be done through your (properly protected) office computer. For example, whenever I download any new software for my DAW, whether it be an update or some new plugin or other tool, I use my office machine to get the files, and then pull the files into my DAW from the office machine. Likewise, if I need to send out anything from my DAW, I can transfer it over the network to my office machine, and then upload to the Net from my office machine.
Selection and Setup of Software
Not only is it important not to burden your DAW with security software, but it is also important to eliminate anything else that is not needed for audio production. This starts with proper configuration of your operating system. Windows is not just one computer program, but actually a big collection of programs all running together. Most of this code is normally invisible to the user. In order to do its work, Windows runs a number of services, each with a specific purpose. At installation time, Microsoft assumes that you will use most of these services, but not all of them are actually necessary for DAW use. It is a good idea to disable the services that you don’t need so that they won’t burden your processor. The most current information on this that I know about can be found on the Saw Studio miscellaneous download page, a document titled “Windows XP Pro Tweaks”.
Along with properly “tweaking” your Operating System, you want to be very selective about the programs that you install on you DAW machine. For the most part, you do not want to install any program that is not directly a part of the audio production process. The obvious “core” application here is the DAW software that you use for recording, editing, and mixing. There are a number of different programs to choose from, at every level of features and at every price point.
Host With The Most?
If you are completely new to recording with computers, and you are not yet exactly certain what you want, you may want to start out with one of the cheaper applications so that you can learn from experience what is important to you. There are even some freeware options like Ardour and there is also shareware such as Reaper. The least expensive “commercial” offering that I know about is called n-track (which sells for less than a hundred dollars).
The opposite extreme in price is represented by various systems that rely on dedicated hardware to do their audio processing. The best known of these, of course, is Pro Tools, which has become pretty much the “industry standard”, so much so that there are many who do not realize that any other system exists. At one time the only way that you could do serious audio mixing and processing work in real time was with hardware DSP, but those days are gone. There are possible reasons for which you may need to have a Pro Tools system, but the ability to do top quality work is not one of them (not to say that PT is bad, just that it isn’t the only game in town any more).
My personal preference leans toward the better “native” systems, which are so called because they use the processor that is “native” to the host computer to do the actual audio processing. My favorite in this category is Saw Studio, but there are several other applications that have a good reputation, including (but not limited to) Nuendo, Cubase, and Sonar. The feature sets and working methods of all of these programs are somewhat different from each other, so it may take some research to find the one that is best suited to your needs (which may not even be one of those listed here).
There is, however, one particular “gotcha” aspect of some of these programs that you may want to consider: copy protection. One of the biggest problems for the creators and sellers of commercial software is piracy. There are always people out there who are ready and willing to steal software rather than pay for it. Many manufacturers try to protect themselves from piracy by using some form of copy protection. There are two difficulties with copy protection, unfortunately. The first is that a determined thief can always find a way to break it, so it is actually impossible to totally prevent theft. The other problem is that copy protection can have the unintended effect of punishing the honest customer. What sometimes happens is that the copy protection suddenly causes the honest user to be unable to use the software that he paid for, and Murphy’s Law pretty much dictates that this will happen in front of a paying client.
In particular, software that requires a “dongle” to be connected to your system for the software to work is absolutely dependent on that dongle. If the dongle fails, (and this does sometimes happen), you cannot use the software at all until you can get a replacement for the dongle. Such a replacement often has to come directly from the manufacturer, and you can be out of business while you wait for the replacement to arrive.
Other copy-protection methods can require a call to the manufacturer to get a new authorization code in case of trouble, and you may be out of luck is the failure happens outside of normal business hours.
The best protection that you can have against copy protection problems is to choose and use software that does not have copy protection. The only top-notch DAW application that I am aware of that does not use copy protection is Saw Studio, but there may be others out there.
Along with this “host” application, you will likely also install a number of “plugins” that add features and power to the host program. Most of the host applications come with a certain amount of processing capability, but the makers of these applications assume that you will want to add some “plugins”, so called because they are the DAW world’s equivalent of the “outboard” gear that would be connected into an old-fashioned mixing console by plugging it in at the patch bay.
When you go looking for plugins, it is important to make sure that they will work with your chosen host application. One “gotcha” here is that plugins are programmed according to any of several different standards. Various companies have developed their own standards for how a plugin should “connect” with a host application. Steinberg’s “VST” standard is probably the most popular, and many DAW hosts will work with it. Microsoft has a standard called “DirectX”, which is also widely supported. Some standards are unique to specific host applications, such as the various plugin standards developed by Digidesign, which for the most part only work with ProTools systems. Another example is the “Saw Native” protocol used with SawStudio, which is designed to allow plugins to make the best possible use of the capabilities of the host program (SawStudio also works with VST and DX plugins).
Another gotcha with plugins is a little thing called “latency”, which is the time it takes for the plugin to do its job. It is easy to forget that a DAW system works by running a collection of computer programs. No matter what any computer program does, it breaks the job up into a gazillion little steps, and the job is always done one step at a time. The only reason that we don’t go crazy with waiting is that the computer often goes through all these steps so fast that we don’t notice the wait. Still, no matter how fast the computer is, it still takes a certain amount of time to finish each job. If a process takes a large enough number of steps, the delay, that is, the latency, can become noticeable.
The host application has to deal with the latency of plugins, usually by applying some kind of latency compensation. For example, you might want to apply a fancy multiband limiter to a particular track of a mix. Suppose this limiter has a fair amount of latency. With that plugin applied only to a single track, that track will actually be delayed compared to all the other tracks. The host application may compensate for this delay by applying a similar delay to all other tracks in the mix.
Some plugins are simple or very efficient, and thus are fast enough that they are said to cause no latency. Strictly speaking, this is not quite true, but the latency is so short that it can be ignored. What is important to remember is that, if you need to keep latency low, such as when you are monitoring live inputs through the host application, you must avoid latency-causing plugins, because using them can cause your DAW to impose a “slap echo” delay on the entire mix.
In trying to tell you what “gotchas” to watch out for when building a DAW, I keep finding more “gotchas”, and there could be many more, but a guy has to stop somewhere, or this article would never end. The idea here, by the way, is not to scare you, but to help you avoid mistakes that could cost you a lot of time and money. Even with all the stuff you need to watch out for, building your own DAW can be a rewarding experience, not only in money saved, but also in terms of really knowing what is going on with your computer. You are a lot more likely to be able to solve problems that may crop up later with a machine you built and configured yourself than with one that was presented to you as a finished package. Of course, it is not possible to warn you of every possible “gotcha”, partly because new ones will keep being invented. Still, I hope the list given here will be useful to you, either in building your own new machine, or better understanding the one you already have.