understanding the PULTEC EQ circuit

[CJ]

0.6 db, so yeah, what i say, the resistor does not mean much. I remember by passing it on the simulator also. If anything, it is a slight attempt to keep the various filter sections from interacting, ie; treble and bass cut.
Or, it could have been a mechanical jumper, just a way to connect the pot to whatever. Seen that done before.

 

[chrissugar]

Chris.

Of course 0.6 dB is not the world but that was measured with the low freq at the lowest value. Probably for the highest low cut position the attenuation is a bit more.
Also don't forget that maybe some people would like to build a precise mastering version and in that case it can be significant. Some mastering EQs with stepped attenuators have 0.5dB or smaller steps and total cut boost around 6dB.
I think it is not an issue for those who build it as a recording or mixing EQ. but it is good to know that for demanding application this 1K resistor can be important.
The original document do not contain that resistor at all:
http://www.purpleaudio.com/pdflib/PultecEQP.pdf

chrissugar

 

[CJ]

Remember, the 1A and the 1R be different.
Will look at the M-ley filter tonite.
cj

 

[CJ]

OK, I scanned the EQP-1A filter section from a brand X equalizer that shall remain nameless.

Viola, no 1 K resistor.

Looks rough on screen but prints nice.

If you want to clean it up and re-scan, knock yourself out.

I can not vouch for the inductor DCR and Q.
The Q numbers do not make sense to me, so take them with a grain o salt.


Right click and save link target as...

Here ya go:

http://vacuumbrain.com/The_Lab/TA/Pultec/eqp_filter.pdf

 

[chrissugar]

Great job Chris,

Now we have an aditional schem and also it is confirmed that the 1K is not in the original.

chrissugar

P.S. Chris, now please send a mail to Pat. :green:

 

[CJ]

Well, Brand X could have left it out.
The only real way to find out would be to peak inside a real Pultec.

Notice that Brand X has changed the 10K resistor to 100K.
Again, if you simulate/listen, not much difference.


Notice Brand X has a few extra freq bands. :thumb:

 

[chrissugar]

[quote author="CJ"]Notice Brand X has a few extra freq bands. :thumb:[/quote]

My mastering version has 12 positions for all the freq switches but that don't change the way the circuit works. I used all the original freq setings and added more intermediate freqs.
I have the feeling that the only diference between the Pultec versions are the makeup gain circuits.

chrissugar

 

[matthias]

@cj:

thanks, for the scan !! do you know what the increase
of the 10k to 100k is good for ??


@ chris:

what about the switches for boost and atten...??
did you use completely lin or lin and log (like the pots in the pultec)

thanks,
mat

 

[chrissugar]

matthias

I made attenuators with the caracteristics of the pots from the original Pultec. For the linear pots I used equal resistors for each step, and for the log pots I used one of those stepped attenuator calculator softwares to emulate the log curve.

chrissugar

 

[CJ]

10k to 100k does not do that much. Maybe a little more output, maybe slightly improved highs, but extremely subtle.
Might have to do with the following transformer. Wait, does brand x even use the innerstage x-former?
That might be the reason. Yes, I remember her saying that they got rid of the "esoteric" x-formers. (excellant marketing-hey, those HS-29's and S-217-D's don't come cheap ya know!)

 

[jhaible]

[quote author="chrissugar"]matthias

I made attenuators with the caracteristics of the pots from the original Pultec. For the linear pots I used equal resistors for each step, and for the log pots I used one of those stepped attenuator calculator softwares to emulate the log curve.

chrissugar[/quote]


I've also been confused from the various versions of these schematics. So I put it in Spice and got this for a "most even" response:

10k lin for high boost
1k reverse log for high cut
10k log for low boost
100k log for low cut

Then, original or not, I don't like a cut function going clockwise.
(read on before you cry heresy!)

I made a (prototype) front panel layout with low cut positioned
left to low boost (same for high cut/boost), zero dB at 12 o'clock position on all knobs, cut going counter-clockwise and boost going clockwise.

This has two benefits:

a) When you're used to work with one knob for cut/boost, the new setup feels very comfortable and familiar.

b) The best part is, that the pots now read like this:

10k lin for high boost
1k log for high cut
10k log for low boost
100k reverse log for low cut

And I could easily get a 100k reverse log pot (but no 1k reverse log).

JH.

 

[CJ]

" TUBES AND POTENTIOMETERS USED IN PULTEC EQUALIZERS

The Types ECC-82 and ECC-83 tubes are equivalent to the
12AU7 and 12AX7 respectively. The manufacturers of these tubes
claim, and our experience confirms, that the ECC series average
substantially lower hum and microphonics than the 12AU7 and 12AX7.


The Low Boost control is Allen-Bradley Type JA-1031. This is
A 10,000 ohm potentiometer with ?Audio? or Logarithmic? taper.

The Low Attenuate control is Allen-Bradley Type JA-1041 or Ohmite
Type CA-1041. This is a 100,000 ohm potentiometer with ?Audio?
or ?Logarithmic? taper.

The High Boost control is Allen-Bradley Type JU-1031 or Ohmite
Type CU-1031. This is a 10,000 ohm potentiometer with
?Linear? taper.

The High Attenuate control is Allen-Bradley Type JU-1021 or
Ohmite Type CU-1021. This is a 1000 ohm potentiometer with
?Linear? taper.

The Bandwidth control is Allen-Bradley Type JU-2521 or Ohmite
Type CU-2521. This is a 2500 ohm potentiometer with ?Linear?
Taper."

 

[matthias]

Hi,


I finished my new pultec eq.... I used switches instead of pots....
everything works, but for low boost and cut I used linear steps...
therefor the "feeling" isn't fine at all...


So I need to get the resistor values to emulate a log curve...

I used this nice calculator:

http://www.quadesl.com/attenuator.html

but I don't know the value for total attenuation (db)
I tried 16db, but the values given aren't useful...

the total resistance for low boost is 10k, but the difference
between 4k and 10k is very low... the same for low cut...

I think to emulate a neg log pot behaviour would be better here??

what do you think ??



If you look at the mentioned schematics...

after the input trafo there's a 620R resistor to gnd

on the x-brand schematic it's a 10k....


does this value depend on the type of the input transformer??
or what is it good for...??

I'm going to use the oep a262a7e as input trafo... it's simply 600:600


will post preview pics soon...

mat

 

[CJ]

decade resistor box-every tweaker should own at least one., invaluable.

You hook it up, then dial in the values, write them down, then your done.

 

[matthias]

I got the final values for the switches...

bass boost 0,6db steps

12 10k
11 5,45k
10 3,57k
9 2,46k
8 1,73k
7 1,25k
6 909r
5 640r
4 440r
3 270r
2 120r
1 gnd

cut 2 db steps

12 75-100k
11 31,4k
10 19,3k
9 13,50k
8 9,67k
7 7,10k
6 5,0k
5 3,4k
4 2,27k
3 1,35k
2 620r
1 0k




but what about that resistor to gnd after the input transformer ???


thanks,
mat

 

[CJ]

Which input x-former?

 

[matthias]

OEP A262A7E

http://www.canford.co.uk/commerce/resources/catdetails/2739.pdf

or Lundahl LL5402

http://www.lundahl.se/pdfs/datash/5402.pdf