"Another Poor Man's" Fairchild 660/670

[rotheu]

Rotheu, I have a few questions...

Are those knobs at the bottom of the unit the DC threshold?

Why did the original unit not have them on the face plate? Or have I missed something?

They control the ratio right?

On the schematic of the original 670 it looks like they are trimmers, is that the case?

Are you feeding the plates of the 6BC8 direct from the 240V supply? IIRC this is over their rated speck. Why not 100V anymore?

Why didn't you use a stepped attenuator for the threshold? Wouldn't it help with stereo tracking?

Thanks in advance...

Cheers
JD

Hi, Matthew

The knobs on the bottom of the unit are the DC Threshold controls. Original unit had them as trim pots inside the chassis and it was factory set for some compression curve. I find it very useful to have access to them at any time, they change the sound of the compression a lot, so I used a stepped switch within most useful range (beyond which the threshold is just way too high) for matching between channels when in stereo mode. IMHO this is more critical than having a stepped AC Threshold control. When stereo linked, both channels are compressed by the same amount no matter what, tiny differences in AC threshold position will not have much effect as long as you set the controls close enough for proper tracking. But it's your choice if you want to use stepped control. Now, with DC Threshold it's a different story, because it changes the ratio and the knee of the compression, so it's a good idea to have them matched.

Yes, I'm feeding 6BC8's with 240V, it's very close to the max. rating of 250V (max. rating of 150V is for Class A1 amplifier at max. current of 22mA per side, here they run at 2mA per side with no gain reduction). I know it will shorten the life of the tubes, but they are cheap and it makes big difference in the amount of gain reduction (extends the remote cutoff point to somewhere around -25V to -30V as opposed to -14V to -18V at 100V), headroom, stability and the sound itself. After many listening tests I chose higher voltage. Again, there is always room to experiment, that's why I was waiting all this time and playing with my prototype, until I actually decided to build it.

 

[Matthew Jacobs]

Thanks Rotheu for explaining that. That's pretty much all the info I needed to re-start work on this comp. Thanks for posting photos of your build, it has really inspired me to carry on working this comp.

JD

 

[jackies]

This is very cool!

 

[haima]

and again, more great and useful information. thanks rotheu!

 

[Kingston]

Excellent documentation. Thank you rotheu.

I especially like the simplicity of the PSU, but I wonder how noisy it might be. It would make sense those high-watt zener shunts add some noise. Have you taken measurements of the PSU?

Also, I wonder if anyone sells a PSU transformer with just the very high VAC output like in this one. I would have use for such a thing. I would rather handle heaters with a completely separate transformer.

 

[KrIVIUM2323]

I'm breathless!
This thread is such a serious diy'ers show! Thank you all sharing with us!

Also, I wonder if anyone sells a PSU transformer with just the very high VAC output like in this one

Kingston, we have in France a E-shop offering custom made toroidal transformers up to 8 secondary (0 to 10000(!!!) VA, 0 to 5000(!!!)V) for fair prices.  Don't know for quality but seems to be ok and good alternative for Edcor's and other overseas manufacturers for power transformers. Lead/manufacturing times seems fast too. The ordering process could be a bit 'complicated' if you do not speak french but if you need translation it's not a problem for me to do (they may have an english page maybe...i do not know?). Here is the link:

http://www.audiophonics.fr/transformateur-torique-sur-mesure-toutes-puissances-p-3675.html

I'm going to order some for projects i'm working on for now and if you are interested could report on them in some weeks.

 

[rotheu]

I especially like the simplicity of the PSU, but I wonder how noisy it might be. It would make sense those high-watt zener shunts add some noise. Have you taken measurements of the PSU?

No, I didn't take any measurements of the PSU, because it's dead quiet. Even with AC filaments and 20dB of gain in the signal amp and speakers cranked up all the way there is just slight hiss. I think the most important when doing point to point is the way you wire the grounds. The way I do it is all my audio input and output grounds are connected to the chassis at the XLR connectors. Power supply ground is connected to the chassis only in one point at the AC inlet. Signal grounds for all amps are connected to that same point. If there is a ground bar, it is connected to that same point and circuits with higher current are always closer to that point than the ones with lower currents. Also keep grid circuitry away from filament wiring and plate circuitry. I always follow this basic rules and never have any hum or noise issues.

 

[MeToo2]

Thanks for sharing your updates. Your build looks really really cool. Love the custom wooden case. 8)

The original layout seems to allow a bit more space than the Hammond box I used, although the signal amp still looks pretty mad.  :

I have a question for you. There's an "in" "out" "balance" switch on your front panel.

In the original 660/670 there was a switch (S101A S101B) that allowed you to put the VU meter alternatively across either the A 1/2 of the winding of the signal amp output transformer or the B 1/2 half or both halves via 2 * 30 ohm resistors so you could measure the current in each half (and thus balance the two halves via the balance pot) or the gain reduction. The disadvantage was that one side of the VU meter had to be at 240V. The Edcor transformer also only has 3 terminals, rather than the 4 of the original, so there's nowhere to put the 30 ohm resistors (R107 R108 of the original).

In your alternative, the VU is wired to ground, which is probably safer and easier.

Now you seem to have added a cunning balance measuring circuit involving a neon indicator, which I freely admit I can't quite figure out.

It seems to attach some of the grids across a pre-biased neon. Is that all 8 grids of one channel via the center tap of the input transformer, or 4 grids of half the channel?
Only thing that would make sense to me would be to connect the center tap of the input transformer and then tune for minimum control voltage, which presumably indicates balance between the halves. I must admit I'd expect to measure something on the anodes (looking at current, or looking for symmetrical voltage swings under signal) rather than the grids (which is basically bias voltage) but maybe I'm not thinking laterally enough.

Do you have an explanation for how your balancing circuit works?

I gave up on this feature and simply measured the balance via an external (high voltage) 2 channel oscilloscope during the build, but if it's an easy retro fit I'd like to add it.

 

[rotheu]

Thanks for sharing your updates. Your build looks really really cool. Love the custom wooden case. 8)

The original layout seems to allow a bit more space than the Hammond box I used, although the signal amp still looks pretty mad.  :

I have a question for you. There's an "in" "out" "balance" switch on your front panel.

In the original 660/670 there was a switch (S101A S101B) that allowed you to put the VU meter alternatively across either the A 1/2 of the winding of the signal amp output transformer or the B 1/2 half or both halves via 2 * 30 ohm resistors so you could measure the current in each half (and thus balance the two halves via the balance pot) or the gain reduction. The disadvantage was that one side of the VU meter had to be at 240V. The Edcor transformer also only has 3 terminals, rather than the 4 of the original, so there's nowhere to put the 30 ohm resistors (R107 R108 of the original).

In your alternative, the VU is wired to ground, which is probably safer and easier.

Now you seem to have added a cunning balance measuring circuit involving a neon indicator, which I freely admit I can't quite figure out.

It seems to attach some of the grids across a pre-biased neon. Is that all 8 grids of one channel via the center tap of the input transformer, or 4 grids of half the channel?
Only thing that would make sense to me would be to connect the center tap of the input transformer and then tune for minimum control voltage, which presumably indicates balance between the halves. I must admit I'd expect to measure something on the anodes (looking at current, or looking for symmetrical voltage swings under signal) rather than the grids (which is basically bias voltage) but maybe I'm not thinking laterally enough.

Do you have an explanation for how your balancing circuit works?

I gave up on this feature and simply measured the balance via an external (high voltage) 2 channel oscilloscope during the build, but if it's an easy retro fit I'd like to add it.

Hi MeToo2,

The output of the balance circuit is connected to the center tap of the secondary of the input transformer. The circuit generates pulses as neon turnes on and off, the 22nF cap and 3.3M resistor set the frequency, which is somewhere between 3 and 5 times per second (Hz). You know, when you compress something really hard and your amp is not balanced, you get the "feed through" control voltage, basically "clicks" and "pops" ? That's how it sounds, but it constant and even. It's being fed to both sides of the PP amp with opposite polarity, just like any hum, noise or distortion, which is supposed to be canceled in the properly balanced PP amp. To balance, you keep your monitors at normal listening level, "unlink" the stereo if linked, put the switch in the BAL position and turn the BAL pot until clicks are "canceled out" or practically not audible. Of course, that's assuming your tubes are reasonably matched. Works really well, simple and fast. Bypass is just some relays putting the unit in "hard" bypass.

BTW, I also used Hammond box, but vertically. It is Hammond 17" x 10" x 3". The rest of the metal work is custom. I wish it had more than 3" depth. I couldn't use all that area behind the meters, because the meters are 2.75" deep. So I had to move the power supply to the other side and add another small enclosure for it, which actually worked well, since it gets really hot and this will allow for better cooling. The SC amp looks kind of crazy, because I tried to keep it as far as possible from the power transformer, so that limited my available space.

 

[haima]

thanks rotheu! that balance circuit seems really cool.

 

[MeToo2]

Hi MeToo2,

The output of the balance circuit is connected to the center tap of the secondary of the input transformer. The circuit generates pulses as neon turnes on and off, the 22nF cap and 3.3M resistor set the frequency, which is somewhere between 3 and 5 times per second (Hz). You know, when you compress something really hard and your amp is not balanced, you get the "feed through" control voltage, basically "clicks" and "pops" ? That's how it sounds, but it constant and even. It's being fed to both sides of the PP amp with opposite polarity, just like any hum, noise or distortion, which is supposed to be canceled in the properly balanced PP amp. To balance, you keep your monitors at normal listening level, "unlink" the stereo if linked, put the switch in the BAL position and turn the BAL pot until clicks are "canceled out" or practically not audible. Of course, that's assuming your tubes are reasonably matched. Works really well, simple and fast. Bypass is just some relays putting the unit in "hard" bypass.

BTW, I also used Hammond box, but vertically. It is Hammond 17" x 10" x 3". The rest of the metal work is custom. I wish it had more than 3" depth. I couldn't use all that area behind the meters, because the meters are 2.75" deep. So I had to move the power supply to the other side and add another small enclosure for it, which actually worked well, since it gets really hot and this will allow for better cooling. The SC amp looks kind of crazy, because I tried to keep it as far as possible from the power transformer, so that limited my available space.

Thanks. Thought it had to be something to do with the imbalance of the control voltage, but I didn't think about the thumps. I hand selected my tubes in sets of 2 channels * 4 pairs * 2 tubes to choose 16 tubes out of a batch of 70. Each individual pair was selected as the closest match possible to cancel each other out. And then I made sets, marking a pair e.g. as channel A, pair 3, side 1 & side 2. So they were pretty well matched to start with. I think the additional screen grid stopper resistors on my (pentode) 6k4p's possibly also helped ease the balancing. Obviously you don't have those on a 6bc8 (triode).

I also like your choice of power supply transformer. I struggled to find anything big enough for the job, especially the 8 amps of heater current. That's why I ended up with two power transformers and two power supplies. Your solution is certainly much more elegant, although it is probably pretty similar in terms of cost because the Hammond transformers can be quite expensive.

I would be tempted to think about adding protection diodes around the TIP50's because there is significant energy /voltage storage in those big capacitors. e.g. a reverse 1n4006/1n4007 diode across emitter / collector for protection at power down in case Ve stays higher than Vc, as well as say a 270V zener across emitter / collector to protect it at switch on (max voltage Vceo is 400V and the reference zener has a delayed switch on and off due to the 10K+100uF). I'd also humbly suggest adding some bleeder resistors e.g. 1megohm directly across the high voltage capacitors so that they slowly discharge when the power is switched off, even under fault conditions like the tubes being removed for service or an HT lead coming loose.

 

[rotheu]

People,

I'm really really sorry, but I have to remove all the schematics I posted. After some listening tests I was given a potential financial proposal and I just can't have it posted. I know this happened on the forum before and I understood the situation, I hope you will do the same now. Whoever has the schematics, you can build it for yourself, but I want to ask you, PLEASE, don't distribute the schematics. I feel very uncomfortable doing this and I hope you will not judge me too harsh.

rotheu

 

[syn]

Rotheu
          great news, and all the best to you and your new product.

 

[ruairioflaherty]

People,

I'm really really sorry, but I have to remove all the schematics I posted. After some listening tests I was given a potential financial proposal and I just can't have it posted. I know this happened on the forum before and I understood the situation, I hope you will do the same now. Whoever has the schematics, you can build it for yourself, but I want to ask you, PLEASE, don't distribute the schematics. I feel very uncomfortable doing this and I hope you will not judge me too harsh.

rotheu

Argghhhhh!!!!

I was reading through the thread and getting more and more excited and hoping one day I would have the time to tackle a beast like this!!  Should saved the schemos last week..

BUT I am even more excited for you that you have such and opportunity, congratulations.  How could anyone judge you harshly?, you like so many here have been more than generous with your time, information and most of all passion.

Thanks for everything and good luck with the new venture,
Ruairi

 

[Matthew Jacobs]

Hi Rotheu

I'm really happy for you. This project always had commercial potential right from the get go. I hope it all works out for you.

Thanks for the permission for those of use that have been following this thread to carry on building this project.

I will respect your wishes and keep those schematics looked away  ;D

For all those who don't have it there are loads of other similar projects available around here and elsewhere.
J

 

[KrIVIUM2323]

I was given a potential financial proposal

8) Yeah! That's a great news!
Don't feel uncomfortable Rotheu: thanks to you and the ones acting in this thread as i now understand how a 660/670 work! Thanks!!!  :-*

 

[ed rees]

Awesome, congrats!

Annnnnd thank god I have this work of genius saved for myself. 

 

[steve honest]

I never try to make my builds look like the units they are based on, but this time I went for original look as close as I could. The original Fairchild 670 just looks too cool not to try to recreate it.

They are some changes to the design, but I'm not sure if I should post them, there was not much interest in this topic. Sorry for crappy pictures, my camera sucks.

Exscuse me if im wrong , or if this is you posting under a psued, but is this not the Eric Valentine 670M? or Unfairchild as he will call them when he builds them for sale?

steve

 

[skal1]

@MeToo2 , i
know it  has been long since you have posted in this thread , i was wonder where you got your pre built , off  the shelf power from for your 660 comp..

cheers
skal1