Another Poor Man's Fairchild 660/670

[rotheu]

Hi, everybody
Had this idea for long time, finally decided to build a prototype and I'm very pleased with what I got. I had chance to use the real thing twice a few years ago and if my memory is good this thingy sounds very-very similar. Right now I'm using it with original timing network but 1uF cap instead of 2uF and it's extremely fast! But I'm planing to have separate attack and release controls for more flexibility. I pretty much used UA175 output amp with reduced feedback for the control amp but kept the front end identical, including the DC Threshold and 6BC8 tubes for signal amp with small changes in the cathode circuitry and lower voltage on the plates. The gain is about 16dB, just like the original 660. You can easily get 20dB of GR and it sounds  tight but at the same time very warm. On fastest position 1 and a lot of GR there is a nice pleasant distortion, which sounds great on the vocals.

[nielsk]

looks very cool, anxious to try it out. Are those just simple 50k log pots input, threshold..)?

[rotheu]

Yes ,right now just simple pots but you can use switches. They are more expensive but better if you build a stereo unit.

[nielsk]

Nice!
how is the balance control working with the VU, is the meters set up to read balance only?
I take it the 100k dc threshold is a trim pot, what would the adjustmant be?

[nielsk]

ie, on the original, the meter is switched to measure plate current and limiting
I take it that you mean C7 (1uF instead of 2 in the timing circuit?)

[rotheu]

how is the balance control working with the VU, is the meters set up to read balance only?

No, the meter only shows the GR. I have another idea for balancing, similar to what EMI did on Altec 436B. It's not on the schematic yet , but I will post it as soon as I try it and if it works

[rotheu]

I take it that you mean C7 (1uF instead of 2 in the timing circuit?)

Yes, that's correct, C7 in the timing circuit

[rotheu]

I take it the 100k dc threshold is a trim pot, what would the adjustment be?

By using 100K in conjunction with Threshold you can have different compression ratios between 2:1 and 30:1. Basically it will change the operation from volume compression to peak limiting. The lower the Threshold and higher the DC Threshold the higher ratio and vice versa.
When the wiper of the 100K pot moves towards cathodes the tube will run in to Class B and will pass only the highest peaks.

[rotheu]

Just adding some pictures of my prototype. Crazy mess! But I love what it does. Can't wait to build it in the chassis.

[skipwave]

That is the coolest thing I've seen in awhile. Makes me a little misty for the old days when cats like CJ and Bluebird used to throw around parts like frisbees.

 

[nielsk]

So what is the deal with the parallel vari mu tubes? I see you have used 4 like the original, the PM660 uses 2, what is the benifit / tradeoff of more vs less? have you tried this with 2 of them instead of 4? would it work with 1?

[Kingston]

So what is the deal with the parallel vari mu tubes? I see you have used 4 like the original, the PM660 uses 2, what is the benifit / tradeoff of more vs less?

more tubes in the sidechain in parallel, more Watts in the sidechain. Those watts translate directly to how fast you can charge the big caps in the time constant network. It's a brute force approach to getting fast, but still clean exponentially shaped attack time, say, below 1ms.

[lolo-m]

  I wish I could do that 

I take it that you mean C7 (1uF instead of 2 in the timing circuit?)

Yes, that's correct, C7 in the timing circuit

If C7 is 1uF, your release times in position 1,2 are two times shorter than the Fairchild... But certainly really helpfull !
Another thing is the 6BC8 grids resistor (220K) can be a problem with some tubes... Some 6BC8 data sheets are saying 500K per grid... If you use 8 of them, you should use 68K in the time constant network... To be honnest, on my PM670 I loaded 4 of them with a 3M9 resistor, one channel was OK and the other was out... Check the voltage on the 220K resistor, if it's not 0 volts, you've got a problem, the resistor have a too high value... Changing it for a lower value will solve the problem but will reduce again your release time...

But again and again, congratulations !!!

[rotheu]

So what is the deal with the parallel vari mu tubes? I see you have used 4 like the original, the PM660 uses 2, what is the benifit / tradeoff of more vs less? have you tried this with 2 of them instead of 4? would it work with 1?

I never got a chance to see PM660 schematic. I tried to read the discussion, but when you don't have the schematic in front of you it's pointless. So I don't know how many tubes it uses. The reason I have 4 tubes is:
 
1)I wanted  to make it close to the original with as less money as possible (6BC8s are only $4.50).
 
2) I wasn't sure less tubes will be able to drive the output properly , especially in deep GR, when they are starving for current.

3) From my experience, the more tubes in parallel you have the easier it is to balance two sides. I guess, on average the differences between two halves  are less for more tubes.

[rotheu]

If C7 is 1uF, your release times in position 1,2 are two times shorter than the Fairchild... But certainly really helpfull !
Another thing is the 6BC8 grids resistor (220K) can be a problem with some tubes... Some 6BC8 data sheets are saying 500K per grid... If you use 8 of them, you should use 68K in the time constant network... To be honnest, on my PM670 I loaded 4 of them with a 3M9 resistor, one channel was OK and the other was out... Check the voltage on the 220K resistor, if it's not 0 volts, you've got a problem, the resistor have a too high value... Changing it for a lower value will solve the problem but will reduce again your release time... Sad

But again and again, congratulations !!!

Thank you! I will try that, but so far I didn't hear any problems. It's very stable

[lolo-m]

So what is the deal with the parallel vari mu tubes? I see you have used 4 like the original, the PM660 uses 2, what is the benifit / tradeoff of more vs less? have you tried this with 2 of them instead of 4? would it work with 1?

The PM660 uses 2 double triodes per channel like the Fairchild.

Edit : You've been faster to answer to the rest of the question ! 

[rotheu]

The PM660 uses 2 double triodes per channel like the Fairchild.

I thought Fairchild used 4 dual triodes per channel???

[lolo-m]

The PM660 uses 2 double triodes per channel like the Fairchild.

I thought Fairchild used 4 dual triodes per channel???

Not you !!!

How's the match 6BC8 versus 6BA6 ?

[rotheu]

How's the match 6BC8 versus 6BA6 ?

I had only 11 6BA6s to try so it was impossible to get 8 of them matched. I couldn't balance them, they were thumping with fast attack/release settings. I used original Fairchild control amp (6V6 output) with them, they cut-ff at around -80V. But honestly, I always liked how 6BC8s compress. I think compared to remote cut-offs like 6386 and 6BA6, 6BC8s are far more aggressive  and "edgy" sounding (maybe because they are on the "edge" of cut-off most of the time ) And the price plays a big role also. tubesandmore.com has new JJ 6386 for $120. F#%k that. I'm staying with 6BC8

[creal]

Garen,
what hell man, i like your proto, youre crazy

here you find more info and schema for the original fairchild.
i hope this can help you to fing the greal
http://audio.kubarth.com/fairchild/