the pimp / mod the Poor Man 660 thread

started this to keep the original support threat a little bit easier.
you are welcome to post your suggestions here.

Ok, here goes....

Would it not be nice to have a MS facility build in just like the original Fairchild 670? Maybe done passively with two 1+1:1+1 transformers. I've some 1517 Lundahls laying around that would be nice for such a thing but I'm sure Edcor has a nice (& cheaper) transformer for the job as well.....

Volker, great idea! :thumb: I have a feeling the PM support thread will be veery long..

Radiance, i was thinking about that as well. I'm planning to try and build up the original 670 lat./vert. switching with some relais around the in+out transformers of the amp section. I'm planning to use Edcors and try to desolder the center tap to have dual windings for that job. Don't know if/how it will work...

I'm also thinking of having the option of either in/out/in+out encoding.

Any thoughts on this? Any other related mods that would add to this, metering, time constants, stereo linking etc?

The original switches the 600 ohm input AND output transformer windings to both encode and decode. So you'd need 4 external transformers for a PM670. Or, as I said in the other thread, hack the appropriate Edcor center taps apart to get separate windings and do it there. Both CT wires come out to the CT lugs. Study the original schematic if in doubt. I have had the Edcors apart in this manner (for repair of mis-wiring from Edcor), but have not tried them in a M/S configuration.

NONE OF THIS HAS BEEN ATTEMPTED IN THE FIELD WITH A PM660!

Matched 1M resistors from attenuator sweepers to center of the two 15K arms would give safety against attenuator failure and open grid condition.

NONE OF THIS HAS BEEN ATTEMPTED IN THE FIELD WITH A PM660!

I'm not going to recommend any of these in particular, only point them out. Feel free to respond with any negatives that alarm you greatly. Note Analag considers the stock stepped attenuators with well matched resistors important.

NONE OF THIS HAS BEEN ATTEMPTED IN THE FIELD WITH A PM660!


Slightly stock attenuator variations

The existing method on the secondary keeps the level in the input iron consistent regardless of setting, and reflects a higher 1800 ohm Z back to the source(s).

If you use a 12 position at 2 dB/step you have a range of 0 to -22 dB, etc etc. Is it enough? Analag says 12 position is fine, more steps is a matter of finer control. Do you want or need finer steps? 24 position at 1 dB/step seems reasonable to me. We are supposed to see 18 dB of gain available, so 22 dB range would get you down to both unity and -4, and there would be 18 dB of make-up gain available.

One possibility is to use fewer steps plus a switchable input pad (to make up for lack of more steps) on the primaries, say 10, 15, or 20 db. Could be a 'U' pad on both audio and side chain amps. Amount dependent on your overall attenuator plan.

Stick with dual stepped atten on the GR amp input secondary as prescribed, but use dual pot on the SC amp. That would do away with the 'ultra-precision balance' factor in the side chain, but not in the audio amp path. Considering the number of limiters that use dual pots in the audio path position, I'm not sure how much difference the average punk rock engineer would notice with them only in the SC amp. The Fairchild uses dual linear pots with taps and paralleled resistances for the threshold; basically impossible to source.

You could use dual pots in the audio amp, like the Gates Sta-level and several other pieces. Not as accurate, but clearly works 'good enough'. Not that you will find a 15K dual audio pot; there's one rub. You could slug 50K dual pots down to 15K with paralled resistances, but the effect on the taper may be too strange.


Transformer primary side attenuator variations

The Fairchild uses 1 dB/step input attenuators on the primary side. Looks like T attenuators, but might be ladders.

If you went with a primary side arrangement you'd probably want to stick with 600 ohm input attenuators, meaning lower 600 ohm load on input source(s), and the level in the input iron would vary with setting. You'd have to work out optimal secondary load resistances; maybe dual 15K resistors, maybe not. Only real benefit is you could use existing Daven 600 ohm ladders, which (at this junction) aren't terribly expensive. BUT, ladders will automatically kill 6 dB of input level, unless you add ANOTHER switch to bypass them at the top of the throw. Unless you already have some magically expensive Daven 600 ohm Ts. Or wire up the NYDave dual knob H pads.


NONE OF THIS HAS BEEN ATTEMPTED IN THE FIELD WITH A PM660!

There's been no mention of stereo linking method to date, and it's not on the PM schematic. The Fairchild simply ties the two control voltages together. Here's some food for thought.

Many limiters use diode isolation to ensure dual mono and stereo operation match in timing. The Fairchild fails to correct for this, for better or worse. As is, in the original design, if you toggle between linked and unlinked the time constants will change because of the nature of paralleling networks. We do not have the complexity of time constant switching found in the Fairchild, so it's not as if I'm suggesting we alter something that already mimics exact original operation. As far as I can tell, the only thing possibly changed by addition of diodes is the relatively minor need to overcome the diode voltage drop through additional drive on the threshold control.

NONE OF THIS HAS BEEN ATTEMPTED IN THE FIELD WITH A PM660!

Here's a comparison of the mono Collins 26U-1 with the stereo 26U-2 for example. This is not to be taken as a cookie-cutter drop-in, given the differences in design. The Collins being the more common type of side chain driven from a high impedance source with fairly high bias voltages present, and the PM660 being a side chain driven from a low impedance source with fairly low bias voltages present. The Collins manual describes the 1N459's as preventing feedback into the 1 mfd caps, to avoid increase in attack time from effective lowering of 1 mfd value.




The unattached connection from the 0.01 cap below runs to the link switch, then to the other identical side chain and audio channel.

Note it's very easy to separate the input of the side chain amp and implement totally separate external side chaining.

Doug,

thanks for all this great suggestions :thumb:
I hope I find the time to build a second unit, which will be the "emrr edition".
got a mail today from analag that he is busy at the moment, but observes this thread (+ the support thread) with interest. I'm sure he will comment when he finds the time.

[quote author="emrr"]Note it's very easy to separate the input of the side chain amp and implement totally separate external side chaining.[/quote]
never thought of this before, but yes, this is easy to do.
this is also a way to switch between "feed backward" and "feed forward" compression. have to try this.

Lots of ideas and thougts here! :grin:

[quote author="emrr"][quote author="[silent:arts]"]this is also a way to switch between "feed backward" and "feed forward" compression. have to try this.[/quote]

Hadn't thought of that![/quote]

To make a feed-forward option how would you connect the SC transformers? Paralell (would give too low input impedance i guess?), series or series with only "half" winding..or do we need some type of buffer from the input?

This is maybe not the first thing i will try but i find it very interesting and educative.

feeding the sidechain transformer in parallel to the input transformer should work fine.

I like this:
NONE OF THIS HAS BEEN ATTEMPTED IN THE FIELD WITH A PM660!
:wink: :green: :green: :green:
could be a cool new signature ...

Ok. Thanks!

Hi,

Is that ok to feed the sidechain input with the input signal taken before the transfo ?

I wondered about a switch connected to a relay to make the job.
Is that the feedback/ feed forward mod?

Thanks

?

[quote author="Gachet"]?[/quote]

[silent:arts] covered it in the previous posts. There are a lot of ways to make the connection, depending on whether you want back/forward and/or side chain input.

NONE OF THIS HAS BEEN ATTEMPTED IN THE FIELD WITH A PM660!

There's been no mention of stereo linking method to date, and it's not on the PM schematic. The Fairchild simply ties the two control voltages together. Here's some food for thought.

Click to expand...

Sorry, can you describe solution for PM? Do you say that in stereo mode PL1 (sec center tap of input tx) have to be short connected from L and R channel? Also, can you draw that part with diode isolation?

[quote author="Moby"]Do you say that in stereo mode PL1 (sec center tap of input tx) have to be short connected from L and R channel[/quote]
Yes (the "original" way without diodes)
http://www.groupdiy.com/index.php?topic=28274&start=356

bridge rectifier diode experiments with moby:
http://www.groupdiy.com/index.php?topic=28274&start=666
(note: starts with post nr 666) :green: