Design tips for a discrete version of Pultec PEQ1A or MEQ5?

[pepe]

I would like to hear people thoughts about the discrete versions of PEQ1A (EQP-1A3) and MEQ5. Are they that good for tracking/mastering compared to other equivalents?
And the best part with this project must be its simplicity, and because of that I can’t stop wonder, why are there so few in this forum who actually have built one? Can a project really be simpler than using turret boards and wire point-to-point?

As for the design,

IN -> Transformer -> PEQ1A Filter -> Discrete Amp (16dB) -> Transformer -> OUT

IN -> Transformer -> MEQ5 Filter -> Discrete Amp (10dB) -> Transformer -> OUT

What components would you recommend?
Some thoughts are Sowter 3603 transformers and John Hardy’s 990C discrete amp.
Or should I drop the transformers and use electronically balanced instead?
As for the inductors, Toko is an alternative as well as the more expensive Sowter/Cinemag. But will I gain that much Q-wise and sound-wise if I chose the later ones?
Capacitors: Orange Drop or Wima?
The parts must not be Pultec correct - my goal is rather to achieve a well designed, good sounding, all around EQ to mess with. I would appreciate any input on this.

 

[Guest]

I count loads of threads and many people building them
It's next on my list of wants..
Try putting
+pultec +neve

into the search

 

[drpat]

...

 

[pepe]

Thanks for the superior advices. I think I will go for the JLM amp then, maybe Forsell’s JFET would be a good choice as well but I already got that in the opto comp.
As for the transformers I’ve been told that it is the input transformers that affect the sound the most, and if so, will I for example get more of a Neve type sound just by changing to Sowter 9165 (Neve P/N 31267)? Ask this since the JLM99V amp seems to be of Neve type. And when I read the specification for Sowter 3603 it says for general purpose, so what makes them so special then. Could I obtain similar result by choosing some cheap OEM’s? Can anyone please give me some hints about “the sound is in the iron” as well?

 

[CJ]

the output for the Pultec is wound on a Nickel core, on a 12 L lam, which is no longer available.
The die is in the dumpster somewhere over near Arnold's old plant near Mojave.

Most folks never wind a tube output on a Ni core, everyone thought EB Harrison was nuts, so we end up with a unique situation, a part used in a new way, giving a new sound.

Add to that the old caps for the bass section and some cool tubes, and you have the funk.

 

[drpat]

...

 

[jrmintz]

HERE is a link to a thread about a solid-state pultec I built a few years ago. It sounds great.

 

[Purusha]

And here is my solid state "Pultec" version with Neve gain stage... sounds awesome BTW :wink:

LINK

 

[syn]

[quote author="pepe"]Thanks for the superior advices. I think I will go for the JLM amp then, maybe Forsell’s JFET would be a good choice as well but I already got that in the opto comp.
As for the transformers I’ve been told that it is the input transformers that affect the sound the most, and if so, will I for example get more of a Neve type sound just by changing to Sowter 9165 (Neve P/N 31267)? Ask this since the JLM99V amp seems to be of Neve type. And when I read the specification for Sowter 3603 it says for general purpose, so what makes them so special then. Could I obtain similar result by choosing some cheap OEM’s? Can anyone please give me some hints about “the sound is in the iron” as well?[/quote]

I had the same info... but...

http://www.auroraaudio.net/dcforum/DCForumID1/302.html
Thank you

Now that i have used 31267, i could say that is a very nice, and in a appropriate circuit it does contribute to the overall sound very nicely...
Further on, output type does change the overall sound, sometimes dramatically .
You will get more Neve sound by using a Neve circuit. JLM Hybrid
driven circuit sounds insanely good, not like a Neve. It depends what you need...
I like Sowter 1166 very much in Neve output stages...
I would go solid state input....
Hope this helps a bit...

 

[pepe]

I believe it would be the way to go for a pure Neve experience, I also believe it sounds amazing. But it seems like a JLM99V should be chosen to gain a somewhat more transparent design, which at least is my first goal.

For a mastering version, any thoughts about 11 discrete steps in the cut/boost/bandwidth section and what values to chose from?
Seems like these should do the trick,
http://www.groupdiy.com/index.php?t...order=asc&highlight=jlm+audio+pultec&start=15
Since I am planning to build a 2-channel version of this thing, I thought it would be nice to add this feature.

 

[pepe]

I’ve done some rough simulations and noticed that by using the new “extended” values for the lower cut & boost section the desirable mid cut effect is gone. Anyone noticed this?
No, there is nothing wrong with those values. But a certain schematic with a 100kOhm resistor on the out causes this behaviour, do use 10kOhm as in Gyrafs version :wink: .
I’ve also concluded that the trick of adding the mid cut filter at the end of the EQP-1A3 section did result in too much interaction. Can anyone confirm this?

 

[pepe]

I’ll take the HF-cut, LF-cut, and LF-boost section from gyraf’s schematic since the cap values seem to be cleverly chosen, and I like the frequencies too. According to my calculations the frequencies are,
HF-cut: 4, 6, 8, 10, 12, 24 kHz. (impedance assumed to be approx 200 ohms)
LF-cut: 20, 30, 40, 60, 90, 130 Hz (impedance assumed to be approx 120 kohms)
LF-boost is: 20, 30, 40, 60, 90, 130 Hz (impedance assumed to be approx 5500 ohms)

As for the HF boost section I’ll use the extended Pultec inductor ordered from ioaudio, but I got really confused when I saw some cap values in the related schematic. I most point out that the cap values looks good but there are some strange ones, such as 36.9nF, 13.3nF. I thought these were very exact cap values for a reason but maybe not?


Inductance-Reactances-Capacitance-Frequency

26mh ----| 2613-3014 | 3.3nF ---- 16kHz
26mh ----| 2287-2419 | 4.7nF ---- 14kHz
26mh ----| 1960-2139 | 6.2nF ---- 12 kHz

39mH ----| 2450-2567 | 6.2nF ---- 10 kHz
39mH ---- 1960-1989 | 10nF ---- 8 kHz

78mH ----| 2450-2393 | 13.3nF ---- 5 kHz
78mH ----| 1960-1989 | 20nF ---- 4 kHz

155mH ----| 2921-2411 | 22nF ---- 3 kHz
155mH ----| 1947-2040 | 39nF ---- 2 kHz

312mH ----| 2940-2875 | 36.9nF ---- 1.5 kHz
312mH ----| 1960-2040 | 78nF ---- 1 kHz

In this design an ideal reactance seems to be around z = 2000 ohms which I assume will give the desirable Q-value. Anyway, 1.5 kHz says z = 3000 ohms but according to simulations this will only give a little narrower Q (not necessarily a bad thing).
Also, every cap value should be as close as possible to the inductive reactance to give the desirable frequency. Therefore, not many cap values are that close, not even 36.9nF, 13.3nF. So where do those values come from?

Can anyone confirm my calculations and results?

 

[CJ]

Thats just calculator dribble.

The problem with the Pultec's is that, given your limited number of freq stops, you may or may not have the perfect box for the track you are working with.
So cap values are not really too critical unless you know exactly what band you want to tweak.

It's hit or miss, so just build it, then if you have to mess with a cap, its a 10 second change out, must faster than three days of reading and typing threads and setting up an excel sheet to run 6.3 F L and 1/6.3 F C all day.
.