PM1000. Extending low end

[JW]

Hey folks,

I did a sweep of my PM1000 preamps the other day, and they're rolling off more low end than I would like. I tried increasing the size of C37, the output cap to upwards of 1000uF, but it's still rolling off. It rolls off low end even if I use the summing buss output instead of the direct out.

I'm using the original transformers, both in and out.  I have a 3.3uF cap in C1

Has anyone else noticed how they roll off the lows?

Any other solutions for this?

 

[Whoops]

Hi,
Do you have the Freq response graph of your sweep?

 

[mitsos]

Hi,
Do you have the Freq response graph of your sweep?

+1

And also a schematic in case any modifications have been done.

 

[JW]

The best copy of the schematic I can find is in the manual. It can be found on pages 10 and 11 here: http://www.pmerecords.com/Docs/PM1000E_2.pdf

C1 (.47uF) has been changed to 3.3uF. In practice, these are actually already 3.3uF tants on most PM1000 channels.

C18 off the fader (.47uF) has been changed to 3.3uF. I THINK I remember these also being 3.3uF tants originally. It's been awhile.

C37 has been replaced with 1000uF. This cap apparently has an effect on the low end, which is why I've gone so high with it.

The original Tamura GA80080 output transformers are wired to the (-) leg of C37.

I am looking for a way to convert the wav file into a frequency response graph (???), because I can only tell by the waveform and what my ears are saying concerning where in the sweep the rolloff starts.

Is there some freeware around for that (for Mac?)

Thanks!

 

[emrr]

I don't even have to look at it to say C37 will have a practical limit, and too high may just mean more DC leakage, which if it's feeding an output transformer may translate to core saturation, less low end.

You need a real measurement for sure.  Ears can not be trusted.  Ears can not be trusted.  Ears can not be trusted. 

A real measurement of the transformers by themselves will give an idea of what's even possible. 

 

[mitsos]

Is there some freeware around for that (for Mac?)

Thanks!

You can try a demo of Fuzzmeasure, they say it's limited, but doesn't specify how.  OTOH, if you're a student or teacher (or know someone who is) you can get the full "personal use" version for $50, which is a great deal.

Agree, 1000uF sounds awfully large.  Are you feeding the busses simultaneously or just direct out?  In my notes I wrote that somewhere on tapeop Jim Williams suggested changing the output transistors to MJE172/182 for more output and less THD, and upping C37 to 330uF for use with a 600:600 transformer. 

But as Doug says, Ears cannot be trusted, so try fuzzmeasure first.

 

[JW]

Okay,

Thanks folks. I'll try that demo. I'm also waiting to use my wife's computer which is a little more up to date and see if the Room Eq Wizard program will be useful http://www.roomeqwizard.com/ if fuzzmeasure doesn't do it.

Anyway, concerning C37, I had a 330uF cap in there for awhile. I've just in the last couple days tried a 1000uF to see if I could get any low end improvement, but it was the same as with 330uF cap.

Mitsos, I've heard a little discussion (on this board somewhere) about a 600:600 transformer being a little less than ideal for the output as it is. But the low end is exactly the same if I come out of the buss, without anything connected to the direct out of the channel (except the transformer) So, to answer your question, yes, the output of C37 is going both directly into an output transformer, and to the buss amps. 

Just to clarify, we're talking about the transistors Q6 and Q9, which are right before C37? (Hardly legible on the schematic.) That would be MJE172 for Q6 (top) and MJE182 for Q9 (bottom?)

 

[mitsos]

I think you're looking at Q8 and Q9, kinda hard to read, but the top one is an NPN, bottom is a PNP.  I think 182 is NPN and 172 if PNP, but make sure to check the pinouts.  I'm going to try BD139 and BD140, cuz I have those around. Anyway, it's the same 4 transistor amp all over the console, and the echo mix amps are the same as the bus summing amps.  See master schem attached. 

Good luck with the testing.

 

[JW]

Okay,

Finally got back to this. Here's what a sweep looks like on fuzzmeasure.

 

[Humner]

yeah, that is quite excessive. My PM1000's drop off maybe 2db by the time it gets to 20 hz.

Is this occurring on only one channel or multiple? Did you recap the eq/highpass section? Looks like the 40hz highpass is engaged. Are you using the original slide switches for EQ section?

 

[JW]

Yeah, I did recap the tants in the high pass section. With the same value though. However, the very next cap after the fader there, is a .47uF (I think that's C18 on the schematic) which I replaced with a 3.3uF. Do you think this could be it?

Yes, I do have the original slider switch, and yes this does happen on every channel. (It's a custom 8X2 mixer that I built out of a bunch of extra channels that I've aquired over the years.

Anyway, I have some transistors on the way to try what Mitsos said concerning those output transistors. I think in the meantime I'll try to sweep it without the output transformer and see what happens.

Humner, are you using the original Tamura output?

 

[JW]

Okay,

I swapped C18 back to .47uF, but that didn't change anything.

I also wired up the direct outs to be transformerless, but that also didn't help.

I guess I'll be replacing the output transistors when they arrive. We'll see how that goes.

 

[JW]

Interesting. Here's a sweep with MJE172 and MJE182 swapped in for the output transistors. It's definitely having an effect. Looks like it's only down about 2-3dB by around 40Hz now, but still not as good as you mentioned Humner, and yours have the stock output transistors?

I haven't listened with my ears yet with these new transistors installed, but I imagine there's a few pretty solid preamps with worse frequency response than this.

Still curious to know if anybody else has been sweeping these and how Humner is getting -2dB at 20Hz . . . .

 

[Humner]

Hold on!

I didn't see the decimal point on the first sweep, you are only at -2db at 20hz. That's basically what I was getting.

I thought it was dropping to 20db by 20 hz, Sorry!

 

[JW]

Ahhhggg,

No, you were right the first time. Between your posts, I actually changed that screenshot because the first one had that dip at 15K that actually had something to do with how I was measuring and not the actual freq. response of the PM1K. So I uploaded another one in the mean time that must have actually been something else. Though it somehow matched exactly what you're describing about your PM1K channels.

Really, I'm not crazy. I don't know how that happened, but anyway, I uploaded the correct freq. graph in it's place (post number 9)
This can be easily seen on the waveform. There is much more rolloff than -2dB at 20K. We're speaking here of the channels that don't have the MJE172/182 output transistors which have a freq. response shown two posts ago.

Anyway, here's a screen shot of the waveform (before replacing the output transistors) The top is my converters for a baseline. The bottom is the PM1000. You can see how it dramatically starts to roll off around 100Hz.

Sorry for the confusion!

 

[mitsos]

Any way you can post the second sweep in a higher resolution?  It is obviously better than the first, but can't really see how much. 

I don't think I've ever seen a sweep of a PM1000 channel,  thanks for doing that. 

keep us posted if you have the same results with other channels.

thanks!

 

[JW]

Yeah,

Here's a little better, just focusing on the lows.

I just tried something. I did a bypass from where the circuit enters the High pass section at the top of the treble pot(R39) to "BI" which is right before C18 AFTER the fader. Strangely the fader still had an effect and I got the same frequency response (as the second graph, with the newer transistors) I didn't disconnect anything else, just added that bypass, so the high pass section and the fader are  running parallel to this bypass, just to be clear.

Hmmmm. Almost seems like the signal is being forced through the 40Hz high pass section for some reason.