A Beginners DIY Audio Roadmap

[Scott ODell]

Hello everyone! I'm an Electrical Engineer that is dusting off my analog electronics knowledge after several years in a software-related job. I recently completed a slightly modified version of the "Weak Joe" optical compressor and am very happy with it. For the first time, I feel like I can find a setting that transforms the audio into the best version of itself, unlike the DAW plugins I've used previously.

After this personal triumph, I've started scoping out future projects to beef up my home recording studio. However, I've come to realize I need the (mostly subjective) advice of more experienced DIYers before I can confidently select which projects to pursue. I'm currently limited by resources- both money, and especially time - and would like to focus on projects that add versatility to my setup, but to save time and money by not focusing on that "last 1%" that people talk about so much. I'm mostly interested in preamps (Neve, API, Tube, FET), EQs (Pultec mostly), and compressors (SSL G-series, 1176, LA-2A).

So now for the real questions I need your advice on:

1) How much do input/output transformers contribute to that "vintage sound" as compared with the electronics between them? For example, say I was in a recording situation where a vintage Neve 1073 would get me 100% of the sound I wanted, but instead I used a vintage Neve with the input and output transformers swapped out with a couple properly implemented THAT Corp. chips. By your subjective estimation, how close to 100% would that be?

2) How do cheap transformers compare to more expensive ones? Is the distortion they produce less pleasing? Or do they produce too much distortion? Using the same Neve 1073 example, if I had a clone with the replica Carnhill transformers, then switched them out with roughly equivalent Edcor transformers, how close to the original would it sound?

3) In general, how much does circuit topology define the sound in relation to the selected amplifying component? From listening to demos of Hamptone preamps, it seems that circuit topology is more important. The Class-A, no-feedback JFET preamp certainly sounded much closer to the character of a tube preamp than a JFET-based OpAmp circuit with feedback. Is this your experience as well?

4) In classic equipment (especitally Pultec EQs, 1176, Fairchild 670), how much of the magic comes from the actual utility (the way it boosts/cuts frequencies or the character of the compression envelope) and how much comes from the gain stages/transformers? I'm particularly interested in what you think of Pultecs (are the filters part of the magic?) and the 1176, because I'm fairly sure equipment like the SSL G-series compressor is all about the utility, as there doesn't seem to be much color added by the gain stages.

5) In your opinion, would it be reasonable to separate the utility and gain stage components of classic equipment so you could mix and match them? For example would it be worth it do build a Pultec with the same passive filters, but then use a transparent OpAmp to provide makeup gain. Then you could also build a line amp based on the original Pultec's tube gain stage and plug it in afterward. Would this type of system or separating EQ filters, compression elements, and colorful gain stages sound pleasing, or would the magic of the classic equipment be lost? Obviously with only this equipment, the example is overly complicated, but if there was a situation that you wanted Pultec filters with API color, you would have the flexibility to use an API line amp instead.

Sorry for the long post  . I know all these questions are extremely subjective and that many of the example are contrived and impossible to answer without me testing them myself. But I'd appreciate any opinions or wild guesses that would help me figure out what projects are worth pursuing and in which order I should pursue them.

Thanks,
Scott

 

[joaquins]

For question one and two, I know you have thinked about it but, it's not expensive nor so time consuming to buy one carnhill, one cheap tx and one THAT chip and try it yourself. I wouldn't assign a percentage number at each component, if you think about it, most, if not all, classic preamps has input transformer. Output transformers you can wind them yourself and isn't much trouble with it. For input is a lot more complicated so there you could want to try it out and see what percentage it's for you each option. Cheap transformers may affect freq response too, not only distortion...

For the third, each one has it's own charecter, I don't think it's so true that class A is closer to tube, it has more THD than op amp based generally speaking, but tube as it's own doesn't mean distortion, there are some clean tube pre amps.

For last two q, if you want to build in this way I'd put the equipment in one box and maybe one or two insert points to split it.

JS

 

[baadc0de]

Hi Scott,

interesting post!

I have built a few projects that can be found here and people are using these day in, day out in a songwriters' studio. I have also modded quite a lot of these after I have built them, so I can offer some insight into these..

1.) input and output transformers to my ears change where the bass resonances of instruments lie and add selective distortion. Sometimes they sound like a multi-band compressor, where certain frequency areas are compressed differently. At least that's what I arrived at, talking with the guys who use my gear and who have heard them through their various transformer swaps (like UTC for Altran, then IC front end..)

2.) It seems that different materials sound different and different core / lamination sizes work out to different effects observed above. So, the bass shift could be somewhere else or the "multiband compression" effect could be more or less pronounced somewhere else. What I do find consistent in more expensive transformers is that they are less suspect to picking up noise from around them. Build quality differs too, usually. As for the sounds, I think it's all about preference - some people like deep woody bass, some like a knocking, punchy one, some like almost emaciated lower midrange resonant bass. Technically, these things are distortions and as such make the unit spec bad. However, they may be just what the engineer / producer / artist want. How do you know if YOU want it? I still don't have an answer to that, so I build as much as I like and I seem to build preference through hearing different things and trying different mods.

3.) All I know is that I like less NFB for my preamps and more NFB for compression. It does make a huge difference. I like oversized output transistors like on TG1 and Neve class A devices.

4.) The EMI RS124 built with Sowter's Altec transformers is to my ears better used (on a mix) as a line device with compression off, instead of in. Just sending signal through the device, but no compression. So sometimes it seems to be the sound more than the action. On the pultec I think since I have heard comparisons between the stock unit and a unit that has everything the same, but a different output amp (Neve A) and the second one sounds slightly better, but the sound is more or less preserved, it could be or probably is more about the action.

5.) Yes. I did that in a few designs. For example the TG1 project has something like this stock, with the "clean" and "iron" modes and that is flipped daily for different sounds. You see in that comp particularly, sometimes it's just the nuking action that needs to be on the drums and sometimes it's the nuking action + carnhill transformers and that together does the sound. YMMV, but it shouldn't be too expensive to experiment, especially if you will be using linear opamp stages as alternatives. It could get expensive if you want a tube stage, a discrete solid state stage and an opamp stage and flip between those. One could also say that chaining different units (like GSSL into an EMI, but with compression off) gives you both the sound and the action, just from different units, they each do what they do best on that particular application, for our users at least.

 

[JohnRoberts]

Speculating about subjective perceptions of various distortions, in response to an undefined stimulus (music) is a wreck on a wreck (opinions about opinions).

The sound of dynamics processors should be dominated by side chain manipulations. When the gain element varies the gain of the music in response to side chain commands, that is multiplying the music by that gain change. Secondary to that the gain element may have some characteristic sound quality, then lastly tube vs solid state, transformer vs direct, etc. but the better the quality product the less difference the secondary details should make.

Recall that back in the day the original design engineers were trying to make these products, flat, linear and quiet, not euphoniously colored .

Good luck...

JR

 

[Harpo]

'..but to save time and money..' vs '..but instead I used a vintage Neve with the input and output transformers swapped out with a couple properly implemented THAT Corp. chips..' might be more expensive than fitting the transformers this circuit was designed for when you add the cost of the needed (different voltage and bipolar) powersupply to the balanced line driver chips (THAT1646/DRV134/SSM2242).
Good luck
Harpo

 

[shabtek]

distortion and spectral distribution for various irons and circuit topology combinations is great for mental masturbation...if you want to build something and use it long enough to get a sense of its usefulness:
start out with outboard power supply, in its own enclosure like 51x or similar to suit your goals. determine a multi-pin connector hat will handle number of DC rails you will likely need
outboard will minimize induced hum and be useful for multiple projects--and allow more room for layout in the chassis

 

[Scott ODell]

Thanks for all the replies!

@joaquins: Looks like I'll need to experiment with transformers to see what I think. I'll probably end up getting some cheap ones for prototyping, then swap them out for more expensive ones once the design is finalized. Then after hearing the results, I can make the decision whether it's worth the extra money.

@baadc0de: Thanks for your response, that's exactly the type of opinion and speculation I was looking for. Again, looks like I'll have to experiment with transformers to form my own opinion.

@JohnRoberts: I know this type of speculation is very unscientific, but there's a very unscientific component of DIY audio. If everyone only cared about reducing noise and distortions, we'd never talk about classic equipment and all focus on building transformerless, opamp-based preamps operated well within their linear range. There is a fairly solid consensus that says Neves sound "in-your-face", that APIs provide a "punchy midrange", and that SSL G-series compressors provide "glue" on a stereo bus. I'm just trying to see if there is a consensus on the sound of the individual components that make up this equipment.

@Harpo: I know, bad example, I was just using it as a thought experiment, not something I would ever attempt. In fact, I think I wouldn't be welcome in some bars if people found out I scrapped the transformers in a vintage Neve.

@shabtek: Don't know if I'm ready for a lunchbox yet, but I'll definitely keep it in mind.