question about PRE's output trim

[compusic]

according to PRE's guide, wind the trims all the way up (clockwise) will allow for maximum signal output (+28 dBu), that's 24db headroom above +4dBu. here I want to know if wind the trims counter-clockwise all the way, then what's the output level? It's not mentioned in the guide.

btw, my PRE's trims have been winded differently for each one when I have access to it for the first time. so I don't know the factory-calibrated position. is it just at the middle position?

[Top Jimmy]

The trim is nothing more than a passive output attenuator. I don't know how much attenuation you'll get when fully counter-clockwise. It does state in the manual that the factory setting for the trim is -6dB to optimize for connecting to a 192 interface.

Here's a methodology for calibrating back to factory if you want to that only requires a volt meter, two TRS cables, and the internal oscillator of the Pre.
Set all channels to line in. Set an identical gain level on all channels with pads off. Turn all trims fully clockwise. Turn on the oscillator and run a TRS cable from the oscillator out to Line in 1. Plug a TRS cable into Send 1 and measure the AC voltage across the tip and ring. Turn the trim down until the voltage is exactly half of what you first measured. Do this for all the remaining channels, i.e. plug the oscillator into the channel, measure the voltage, trim it to half.

Don't worry if the voltage measurement on each channel isn't identical. Just adjust the trims so that the voltage is halved.

[compusic]

thanks for reply. your method is one of the working methods.

well, let me explain it in details. maybe Digidesign's technician can make it clear easily.

First I just made a mistake before. I thought the PRE's trims positions (the appearance I mean) shoude be the same. In fact, they do not need to be the same. and I just have tested and can be sure that they are at factory-set positions, that is –6 dBu below its maximum output of +28 dBu (+22 dBu), and that's for output to 192 I/O.

But my interface is 96 I/O, and 96's headroom is 14db above +4 dBu, that means the maximum level 96 I/O can accept is 18 dBu. here comes a problem. since the maximum level PRE can output now is 22dBu. and I can see that even the PRE's input peak meter is far from clip, Pro Tools shows that the 96 I/O's input is overloaded. btw, both 96 and PRE work in +4dBu mode.

then I let 96 I/O output a 1k sine signal with level of -6db fs, and feed it to PRE's input then output to another input of 96 I/O, and check that 96 I/O's input level with a software meter. here the 96's input setting is linein, 0 db gain, and it seems the impedance doesn't matter here. at the trims factory-setting positions, the sinal 96 receives is -12 db fs. It seems PRE add the -6 db attenuation to the linein level and then output it, right? now I wind the Pre's trim to let it -10 dBu below its maximum output by checking the software meter to make sure it shows -16 db fs, here 28-10=18, that's exactly the maximum level 96 I/O can accept.

Do I set it correctly? if so, here comes another problem, the same line level when enters 96 I/O directly and enters Pre first then into 96I/O has different recorded digital level, if want to be the same, I have to set Pre's gain to 10 db, in fact 9 db since it's the closest value. Is that the only way to let the sinal have the same recorded level?

and btw, even I make settings like that, here the Pre's output level and 96I/O's input level are compatible, I still can see that when PRE's input peak level meter shows -1 or -2db, Pro Tools shows that 96 I/O's input is overloaded.

what a strange behavior!!!!!!!!!!

and btw, I also have done a loop test, that's 96 I/O's output directly to it's input, the output level and input level have the same value. It shoude be like that. just make sure 96 I/O works correctly.

[Top Jimmy]

Your method for trimming the ouput of the Pre for the 96 I/O is adequate. You have also rightly discovered that you have to apply 10 dB of gain to a signal to match the output. The trim works as an output attenuator. A signal that comes into the Pre at +4 dBu will be trimmed to -6 dBu with the gain at 0. What this accomplishes for you is maximizing your signal/noise ratio. If you're seeing the over indicator in Pro Tools before you see the clip indicator on the Pre, then your settings are good. I believe it to be preferable to everload the A/D converter before the circuit clips.

[compusic]

If you're seeing the over indicator in Pro Tools before you see the clip indicator on the Pre, then your settings are good. I believe it to be preferable to everload the A/D converter before the circuit clips.

thanks for reply. but there's one point for that I don't agree.

I think it's preferable to overload PRE before 96 I/O gets overloaded. since I only need to keep my eyes on PRE only. if PRE is not overloaded, then 96 I/O can't be either. if ever there's clip, it should be on the PRE side, and as far as I set the gain to not let PRE overloaded, I'm sure 96 I/O is not overloaded.

according to your prefer,I have to keep my eyes on both PRE and 96 I/O, since if PRE is not overloaded, but there's possibility that 96 I/O is.

[Top Jimmy]

I understand your reasoning, but I submit the following for your consideration.

In all analog gain stages, the last 6 dB of gain has the greatest amount of distortion. It's in the upper levels of gain output where the "harshness" or "hardness" lives. By aligning the clip point of the Pre with the clip point of the A/D converter, you maximize the signal to noise ratio, but have moved the circuit into a gain range where its performance may not be as ideal as it could be. Personally, I would trim the Pre outs by either 4 or 7 dB (for 6 or 3 dB headroom respectively) instead of 10. Why? Because it would prevent the circuit from ever reaching its maximum gain since you'd be clipping the 96 I/O.

[compusic]

thanks. but, according to your words, the distortion of the last 6 dB of gain occurs only in 96 I/O's circuit, not in PRE's circuit?

[Top Jimmy]

That would only happen if you rely on the clip indicator on the Pre. Instead, you should rely on the over indicator on the 96 I/O, or in Pro Tools.

Look at it this way. We have an API 3124+. It clips at 30 dBm. If I brought it over to your studio, would we run the output at 30 dBm? No, because it would overload the A/D converter in the 96 I/O. We would adjust the gain on the 3124+ so that it doesn't overload the 96 I/O. Yet, the 3124+ still has 12 dB of gain that never would get used. Is that a waste? Not at all. This unused headroom is part of the key to the quality of the API.

Optimum gain structure is where all gain stages are aligned to clip at the same level, and that level would also be aligned to digital zero. If having headroom available in the preamp is good for the sound quality, why would we want to trim the outputs so that the preamp clips at digital zero? The only thing optimum gain structure gets you is maximum signal/noise ratio. It does nothing good for the sound quality.

So I put this to you. If you care for sound quality over signal/noise ratio, you would not trim the Pre outs by 10 dB. Instead, you would choose a trim level that leaves some unused headroom in the preamp.

[compusic]

well as my understanding of your words, you prefer to let preamp (here is PRE) have more headroom than let interface (96 I/O) have, for the purpose of sound quality.

but why not let 96 I/O have more headroom? does the rule - more headroom better sound not apply to 96 I/O?

and if I set both of them to clip at the same level, I can keep the input level low if I want to both of them have much headroom, no need to worry about that whether one has much headroom and the other doesn't have enough.