UA175 inspired variable Mu compressor

I've asked a lot of questions about this in the past, so I'll keep you guys updated on the progress of my JH175 project.

I've made same first listening tests with the still unfinished compressor. These sound promising, so it's time to start publishing the stuff.

No sound samples etc. yet. This will take some time.

At the moment, I have pictures of the construction, and a PDF of the power supply.

You can find the files here:
http://www.oldcrows.net/~jhaible/vari_mu_compressor/

I think the power supply is a little unusual in two ways:

The DC heater voltage regulator is a very-low-drop design, using a MOSFET as series regulator, and a voltage multiplier to create an auxiliary voltage to control the MOSFET's gate voltage.

The anode voltage uses two torioridal transformers back-to-back (nothing special so far), but the second one is connected as an autotransformer, thus providing more than 230V AC.

I've also tried to preserve the original PSU's behaviour without using a choke, and using an electronic inductor instead.

JH.

Looks FUNKY, man!

Did you do any mods to the audio schematics? Are you using the original 6BC8 for gain reduction?

:shock:
great!
:thumb:

[quote author="rafafredd"]Looks FUNKY, man!

Did you do any mods to the audio schematics? Are you using the original 6BC8 for gain reduction?[/quote]

I've added a trimmer in the cathodes of the BH7A to balance the current. Apparently the original transformer had a gap (was used in SE circuits!), and my Sowter (Fairchild signal amp output transformer) didn't like the 1.5mA unbalanced DC current too much.

6BC8 - yes. I only tried one, and it worked. I had to set the balance trimmers to quite extreme positions for CV feedthru rejection, though.

JH.

Tonight I ran some music thru this, with mixed results.

A readily mixed CD (I played some weather report, and over-comressed it) sounds great thru the vari mu limiter.

My Wurlitzer epiano - with high dynamic range, and with some keys being a lot louder than the others - did not sound so well. My LA2 clone sounds much better there than my 175 clone. Too much (unpleasant) distortion.

Then I tried a guitar sound from the (physical modelling synthesizer) VL7. Here the vari mu was great as well.

Preliminary conclusions: The vari Mu doesn't like too much of dynamic range. Otherwise you can easily overdrive the grid of the first tube without that tube having a chance to decrease the signal.

Maybe this is why variable mu compressors are favored as mastering tools, to just give the final touch of a few dBs of compression?

JH.

I just worked on a Fairchild 670 this week.

Just want to say that thing can really nail a lot of gain reduction onto a signal without buzzing it up much at all!

...uses 8 x very hard-to-find tubes for the GR though...

Keith

I had a 175 a few years ago. It had no problems over a wide gain reduction range. I used to peg the meter on bass, guitars and on vocals. Smooth warm fat etc etc etc. Like an 1176 SHOULD sound. But valvier(Yuk! did i say that?). i had loads of problems with it tho. it was very un-reliable. mostly tube related. the vari-mu tube would need replacing all the time! seemed like no more than 6 months or so. I had to adjust it to stop thumping so often that i did it before every mix as a precaution! Small price to pay for such beauty IMHO. I am sure you can get a wider range. Enjoy! Wish i still had mine . . . . . Sigh . . . .

Keith, lucky man to be handling a 670. It's been a very long time indeed for me. Love m+s devices to bits, me . . .

Andy P

[quote author="strangeandbouncy"]I had a 175 a few years ago. It had no problems over a wide gain reduction range.[/quote]

Andy - can you quantify "wide"? 20dB? 40dB?
20dB GR or even a little more seems to be no problem on mine, but I my optoelectronic compressors (LA2 wannabe and a little DIY circuit with SSM2017 and VTL5C4) have a much larger range.

JH. - trying to understand if it's basicaly how it should be already, or needs furher tweaking.

Problem (partially) solved.

The signal path is fine. Actually, with compression turned off, but the
signals running thru all the tubes and transformers, the Wurlie sounds wonderful - warm and fat.

The distortion I heard was not from overdrive, but from the time constants apparently. Turning attack time and release time full up makes most of the distortion go away. Always speaking of the Wurlitzer epiano going thru the compressor - playing full-handed chords apparently created nasty intermodulation when the vari mu tubes were modulated. Classic compressor time constant issues, ONLY that it seems more severe than ordinary. Still a little strange, but the fun working on this project is clearly coming back!

JH.

how do I quantify "wide"?

well, - a vu meter only reads down to 20dB, more than I'd USUALLY use . . .

You shouldn't find too much of a problem related to attack/release if I remember rightly, Mine was very forgiving . . .

Andy P

Short report from the weekend:

The distortion problem with the Wurlitzer *is* because of the rectifier time constants. I've increased the release time constant to 1s (10Meg resistor), and then, together with the longest attack time, the problem is mostly gone. (but 1s is way too long for practical use.) Then I checked with my 1178 clone, and there I get a similar distortion, jut not quite as severe. The only one of my compressors that works without any of these effects is the LA2 clone. I guess the dynamic time constants of the LDR are just what an e-piano needs. I also tried with a Rhodes (not the real thing; a ROMpler this time), and its similar. With a grand piano, there is no problem.

Apparently electric pianos are very unforgiving to that kind of distortion. Mellow tones (not much of harmonic content in the Sustain phase of the tone), plus rather hefty transients when a whole chord is played loudly, plus polyphony that allows a lot of notes to carry a lot of energy.

Good news for me: As my 1178 clone shows a similar problem with that kind of input material, and as it's working nicely for years to limit my stereo mix, I'm gaining confidence that my 175 project isn't faulty either. Just not the first choice for e-pianos.

JH.

[quote author="rafafredd"]Looks FUNKY, man!

Did you do any mods to the audio schematics? Are you using the original 6BC8 for gain reduction?[/quote]

Not yet, apart from the transformers and the bias adjustment. But I have an idea for a future modification:

There is this other compressor, UA177 I think, which has variable compression rate. As the signal amp and control amp is one and the same thing, they use a multitapped output transformer in order to feed a reduced level to the rectifier for the lower compression ratios.
I thought about how to emulate this.
I just have to attenuate the signal before it goes into the output amp (after the interstage transformer), and then compensate for the loss in the output attenuator section. Instead of a single output attenuator, there will be two in series, both bridged T. The second one has 2dB steps (as before, but without the 6dB loss), and the first one has 3 positions with just the right attenuation to compensate for the attenuation before the amp. Of course the overall output level will be lower: That scheme eats 10dB in order to get a range of 4:1, 8:1 and 12:1, but with the 6dB recovered from using a bridget T instead of a Ladder, thats just 4dB less than before. That means, threshold of limiting (in 12:1 mode) is +9dBm maximum. Enough for my application. I'll post details when I have built it.

JH.

[quote author="SSLtech"]I just worked on a Fairchild 670 this week.

Just want to say that thing can really nail a lot of gain reduction onto a signal without buzzing it up much at all!

...uses 8 x very hard-to-find tubes for the GR though...

Keith[/quote]

Now that 670 is a different machine, but as the working principle is much the same, I'm taking this hint very seriously.

So I'm asking, are you talking about a lot of gain reduction on an epiano (played solo, complete with bass notes and chords, mellow tone) ?

JH.

Try different tubes, try to get a closer match. One thing to try (and this is how I balanced my 660 clone) was to feed DC into the control voltage port, and measure the current on each triode throughout the range - maybe ten or so points. The distortion you are hearing at short attack times is probably what I heard when I had unmatched tubes and just moved the bias pot to equalize the currents. That goes away if the two sides track properly. CV feedthrough can change at different gain reduction points.

[quote author="dale116dot7"]Try different tubes, try to get a closer match. One thing to try (and this is how I balanced my 660 clone) was to feed DC into the control voltage port, and measure the current on each triode throughout the range - maybe ten or so points. The distortion you are hearing at short attack times is probably what I heard when I had unmatched tubes and just moved the bias pot to equalize the currents. That goes away if the two sides track properly. CV feedthrough can change at different gain reduction points.[/quote]

Excellent point. I have trimmed the CM way down, but that was at one point only, and the trimpots were almost at one end position, which may indicate a lot of internal unsymmetry forced into place from the outside. I'll try different tubes - fortunately I've bought a couple of these.

JH.

[quote author="dale116dot7"]Try different tubes, try to get a closer match. [...] The distortion you are hearing at short attack times is probably what I heard when I had unmatched tubes and just moved the bias pot to equalize the currents. That goes away if the two sides track properly. CV feedthrough can change at different gain reduction points.[/quote]

You were so right!

I tried all 4 tubes that I have, and what a difference!

It's only, that two of them are far *worse* than the 1st one. 8-(

At least the 4th is slightly better than the 1st. I'll leave this one in the circuit for now. I'v also ordered a couple of matched/balanced 6BC8's today - let's see if these will be better.

Quick audio test: Sounds like #4 is a little better.

JH.