PM670 raspberry shake flavour

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Hi Mike,
Can you give us more details about your mods?
i like to know is you use the original time constants of farchild.

i very intresting to add  the mid-side, feedback and sidechain highpass in my PM670.
Mabeby you can post moore info to help us to trun our Poorman to Richman.
Thx
Creal

 
I'm afraid I still lack the skills/webspace/will to post audioclips. Let's just leave it to the statement that it sounds like it looks.

Someone asked about cost. It'll be less - maybe one half or a third - compared to anything remotely equivalent on the open market, but the amount of time wasted on any build always takes away any such benefits. On this one I estimate 30-60k euros for all the hours at a rate similar to my real job. But I've learned an equal amount, where does that leave me? Counting money and DIY don't mix.

radiance said:
I might go the relay way as well. Did it make a difference "hum wise" compared to rotary switches & unshielded wire?

Duct is also very cool. My power switch is located at the same location but my audio wires are a little further away from it. So far I've had not much hum.
Still...I might go for the duct anyway :)

didn't compare relays vs. plain wires. I just designed a prototype/testbed PCB with enough options so I could start using it right away. building stuff twice is not my idea of fun. I use relays by default because I'm lazy. They guarantee a short signal path, and the DC voltages needed for them will never disturb circuits around them. Also look at the amount of wires already running in this build, would you feel safe dangling more audio around there. I did the duct "by default" for much similar reasons. I measured the benefits in one past project, and until someone hands me a better solution I will just keep doing this.

Ptownkid said:
Did you happen to notice that you are missing a w on the right hand slow?

Yes. Guess if I will ever use this sticker guy again. I thought it was pretty "hilarious" when the same thing happened with my CNC-engraved dual 1176 front panel, by a company not even remotely related to the sticker guy. One of the four perfectly equal dials is missing the number 30. "why" is a perfectly valid question considering I sent them both pure vector DXF format with no layering tricks.

creal said:
i like to know is you use the original time constants of farchild.

No way! Most Fairchild 670 time constants are completely pointless in this day and age. There is no material I know that needs a release time of 20 (or even 5) seconds. The meter just sits buried and doesn't move, nothing happening. We have faders for that.

The time constants are a mix of stuff I gathered from the forum over the years for various vari-mu projects. There were caps ranging from 1-15uF and parallel resistance from something like 50k to "off". For the parallel release used on some of the rotary steps I don't remember what I ended up using anymore, but they are quite subtle. Plain RC covers nearly all uses, so I just provided plenty of range. Most of the network caps are buried under those two big aluminum shell PIO caps in the pictures. For stereo link I used some diode protected configuration.

I didn't document many of the things well. I have a bad habit of just drawing a PCB from misc. notes I have that suits "all options" for the task, and then experimenting while I'm wiring it in. As such I have no good documentation for any of this, except that "it's all here on the forum already".

There will also never be PCB's for sale. They are specialised for the parts I have around, no use to any of you and might as well have done it p2p in the first place since it's only going to be used once. Lukas just makes such good sturdy PCB's that I would much rather do that than p2p.
 
Unit looks great. 

Feedforward versus feedback should have microscopically faster attack time and feel different in nature.  I haven't yet tried it with this type unit to comment realistically.

Kingston said:
creal said:
i like to know is you use the original time constants of farchild.

No way! Most Fairchild 670 time constants are completely pointless in this day and age. There is no material I know that needs a release time of 20 (or even 5) seconds. The meter just sits buried and doesn't move, nothing happening.

Maybe for your envisioned usage.  Classical music needs long release, as does broadcast AGC.  Music with varying movements and dynamically different interludes or breaks require momentary long release.  I rarely use very long release times either, but am very glad to have them when needed. 
 
emrr said:
Feedforward versus feedback should have microscopically faster attack time and feel different in nature.  I haven't yet tried it with this type unit to comment realistically.

I'm not sure I implemented feedforward correctly. It sounds very odd at high reduction, and then abruptly turns too subtle. I think what I have at that setting is actually a hybrid mode of feedforward+feedback.

It's one of the places that will go through a troubleshoot and re-think phase when I fix the mid-side configuration (it's on the same PCB). At least the SC high-pass works really well.
 
Yes indeed I just checked and I have implemented feed-forward incorrectly.

The way it should be is that main input to the unit goes to both vari-mu amp input transformer and the sidechain input transformer but vari-mu amp output is no longer connected to the sidechain at all. No loop-through in the signal chain.

That way sidechain only ever amplifies unaltered input signal, and that (rectified and time constant filtered) moves the grids for compression. Should make for a pretty subtle and natural compression performance.

The implementation I have right now is just wrong, no more comments on that.

I guess, moving further in flexibility, one could actually alter the ratio of feed-forward vs. feed-back with passive mixing right before the SC input transformer. A cross-fader between the main input signal and vari-mu amp output signal. Hey, haven't seen that one implemented anywhere yet! The SC amp certainly has make up gain to "waste" for it.
 
Nice Modded build.

If you want the M/S relay based design I think I posted it somewhere on the main thread when I did mine. I have an eagle file somewhere for it just throw me an email. It works great for me.



Chuck
 
Very nice!
Did you go with the original tubes? Did you have to match them carefully?

Built mine a while ago, haven't done any mods yet. Just getting it going was enough at the time...
 
That way sidechain only ever amplifies unaltered input signal, and that (rectified and time constant filtered) moves the grids for compression. Should make for a pretty subtle and natural compression performance.
I once tried feedforward on a prr varimu built (modified for pcc189 tubes but prrs sidechain)and the result was anything but subtle. The compression was higher ratio, it sounded kind of "jumpy" and with a lot of gain reduction I had negative ratio (more input resulted in less output).  It sounded pretty interesting untill like 6 or 7 dB of gainreduction but then it got weird. It needs some careful implementation for sure. Gyrafs varimu works with feedforward, I think, and it obviously works well.
Feedback is correcting itself in that the sidechain gets less signal with more compression and that way it never goes into negative ratio.
Feedforward versus feedback should have microscopically faster attack time and feel different in nature.
It sounded for me, that FF was even a little bit slower on the attack than FB, but I never measured.
FF is an interesting option, but I found FB to be more flexible in my varimu, as I like working with a varimu in deep gainreduction.
Tobias
 
ChuckD said:
If you want the M/S relay based design I think I posted it somewhere on the main thread when I did mine. I have an eagle file somewhere for it just throw me an email. It works great for me.

You posted that in the original thread already. The way I did it was a slightly different take, just failed the implementation. I chose the wrong output transformer for the task. Works perfect in plain dual mode at least.

mrclunk said:
Did you go with the original tubes? Did you have to match them carefully?

No, didn't go with 6BC8's, these are the russian alternatives. I actually even tested certain russian pentode connected triodes with this 2X9-pin sockets to 4X7-pin sockets adapter I made, but decided I will save them for some future vari-mu build. Forgot to take pictures of those. They look pretty.. uhh.. interesting.

Yes, you have to match the grid curves really carefully.

All the info and alternative tube tests are buried in the original build thread.

hop.sing said:
with a lot of gain reduction I had negative ratio (more input resulted in less output).  It sounded pretty interesting untill like 6 or 7 dB of gainreduction but then it got weird.

I bumped into negative ratio as well, even with my hybrid (failed) implementation. Useful as an effect I suppose.
 
Kingston said:
Kingston-PM670-FP1.jpg
What a beauty!
This is really a unit I'd love to have. Could leave m/s, could be mono,
but man. Vari-mu and flexible times :)
How fast is it? Fast enough for acoustic guitars? For DI electric guitars? Tambourine?
Or do you see it more as a masterbus-compressor/ leveler?
 
Back
Top