Poor Man's Pultec EQP1-A Build Support Thread

[salomonander]

great thanks mate. ill do this tomorrow. i can run it as high as 192khz no prob. not certain about impedance. im running a burl mothership. i assume its low on the output and high impedance on the line ins.

 

[ruffrecords]

great thanks mate. ill do this tomorrow. i can run it as high as 192khz no prob. not certain about impedance. im running a burl mothership. i assume its low on the output and high impedance on the line ins.

If you add a 4K7 resistor across the input it will be reasonably close to the conditions in the EQ. It is beginning to look like this is a common property of inductors of this type of construction.

Cheers

Ian

 

[salomonander]

hey Ian,
you mean between my hot and cold pins on the input xlrs?

 

[salomonander]

alright. so i did not have a 4k7 resistor handy. plus i could not get the rightmark to install on my system. so i just used some whitenoise and an analyzer. im haveing the exact same issues (9042). sorry for no graphs  - but they would look exactly the same. are these inductors faulty?

what are we going to do about this? any possible solution except using a different inductor? gosh i bought lots of them and most likely wont be able to return them i do work at 96k a lot so i cant just ignore this...

ps: on my redd eq - which uses a carnhill as well for the10k boost - everything works fine.

ps2: i just wrote an email to martin at carnhill pointing this out. ill let you know what he says.

 

[ruffrecords]

hey Ian,
you mean between my hot and cold pins on the input xlrs?

Hot and cold of the output XLR.

Cheers

Ian

 

[ruffrecords]

alright. so i did not have a 4k7 resistor handy. plus i could not get the rightmark to install on my system. so i just used some whitenoise and an analyzer. im haveing the exact same issues (9042). sorry for no graphs  - but they would look exactly the same. are these inductors faulty?

It is beginning to look like it is systematic. All inductors have a self resonance but I had not expected it to occur so close to the audio band.

what are we going to do about this? any possible solution except using a different inductor? gosh i bought lots of them and most likely wont be able to return them i do work at 96k a lot so i cant just ignore this...

Do you really have any content above 20KHz?

ps: on my redd eq - which uses a carnhill as well for the10k boost - everything works fine.

The REDD EQ uses a VTB9044  which has a maximum inductance of 200mH compared with 1000mH of the VTB9042. This means there are less turns on the VTB9044 which means less self capacitance which means a higher self resonance which means no problem at 10K.

So the obvious way to fix it is to redesign it for a different inductor like the 9044 which is a pain for both you and me.

ps2: i just wrote an email to martin at carnhill pointing this out. ill let you know what he says.

I have written to Colin at Audio Maintenance and asked him also to contact Carnhill. I do wish manufacturers of inductive components, both inductors and transformers, would publish comprehensive specifications.

Cheers

ian

 

[salomonander]

hey Ian,

yes, its the same on all of my inductors. i did measure some other randome inductors i had lying around and they showed similar behaviour.
as you have said, the only proper way to fix this is most likely to re-calculate for another lower value  inductor. while a pain for all, i think it would be worth it.  if you dont have the time i might try myself but im horrible at math but building an eq that i intend to use for high end bossing  that has this behaviour is not an option for me.

 

[salomonander]

ps: im really no expert but it seems that the oscillation goes further down when using a higher tap. meaning that the issue might only occur at say 30k when boosting 12k&16k but seems to moves down to the audible band when using the higher taps.
this comes from a noob that tested the inductor outside of the circuit but measuring the inductor on itself it oscillated at 18k when using the highest tap.

but i cant really believe that so many people build this without anyone experiencing issues..  did people not measure their units? maybe some builders can help by measuring their finished units?

it also doesnt make sense that the toko inductor that jonas measured, did fine at 100mhz. otherwise id say that there is some misdesign happening (with all due respect) but if the other one works, we might simply have to find an inductor that does the job.
are we having the same issue on the mid band?

 

[salomonander]

what are we going to do about this? any possible solution except using a different inductor? gosh i bought lots of them and most likely wont be able to return them i do work at 96k a lot so i cant just ignore this...

Do you really have any content above 20KHz?

im afraid i do. otherwise no one would record 96k . plus my mastering engineer will ask me whats wrong  up there its no option. i do own around  50 analog eqs and none of them show any behaviour like this. this needs fixing asap. while it might be a minor issue to some, it is desaster to me.  i will try some axial single tap inductors tomorrow.  but how can the most built eq on this forum not even work properly? whats happening here?

 

[jonasmaks]

Hi salomonander,

are we having the same issue on the mid band?

If you would look at the picture in my Reply #835 you see that we have even bigger problem on Mid Boost, because VTB9050 have lower self resonance than VTB9042.

but i cant really believe that so many people build this without anyone experiencing issues.  did people not measure their units? maybe some builders can help by measuring their finished units?

I think people judge by the ear as much as by numbers (measurements) - and it's hard to judge about those frequencies by just auditioning them - you should have at least descent equipment (and be quite young - I probably do not hear as much over 10K as I did 25 years ago). And maybe someone could find those filters artifacts pleasing - without knowing the math and looking at the graphs…

what are we going to do about this? any possible solution except using a different inductor? gosh i bought lots of them and most likely wont be able to return them

Yesterday I have wound my own multitap inductor to mimic VTB9042 - it exibits very similar low self resonance. I will try another core material/turns count during the weekend to see if we could achieve lower capacitance for same inductance.

it also doesnt make sense that the toko inductor that jonas measured, did fine at 100mhz.

I've measured 100 mH toko (or to be exact - murata R1900 series) inductor which did fine up to 40 kHz.
You can find measurements of G-Pultec based on those standalone inductors in my Reply #835.

 

[salomonander]

Hi salomonander,

are we having the same issue on the mid band?

If you would look at the picture in my Reply #835 you see that we have even bigger problem on Mid Boost, because VTB9050 have lower self resonance than VTB9042.

but i cant really believe that so many people build this without anyone experiencing issues.  did people not measure their units? maybe some builders can help by measuring their finished units?

I think people judge by the ear as much as by numbers (measurements) - and it's hard to judge about those frequencies by just auditioning them - you should have at least descent equipment (and be quite young - I probably do not hear as much over 10K as I did 25 years ago). And maybe someone could find those filters artifacts pleasing - without knowing the math and looking at the graphs…

what are we going to do about this? any possible solution except using a different inductor? gosh i bought lots of them and most likely wont be able to return them

Yesterday I have wound my own multitap inductor to mimic VTB9042 - it exibits very similar low self resonance. I will try another core material/turns count during the weekend to see if we could achieve lower capacitance for same inductance.

it also doesnt make sense that the toko inductor that jonas measured, did fine at 100mhz.

I've measured 100 mH toko (or to be exact - murata R1900 series) inductor which did fine up to 40 kHz.
You can find measurements of G-Pultec based on those standalone inductors in my Reply #835.

thanks jonas,

maybe the muratas are the solution then. the osciallation certainly is not.  its totally off. i will check my surplus dealer tomorrow lookng for some 100mhz inductors just to get an idea.

 

[salomonander]

just looking at the mid boost now '835. thats totally off and broken. its clearly in the measurable region even with the worst test equipment.  omg.  but the toko looks fine. the bell is even.  darn carnhills...

 

[ruffrecords]

hey Ian,

yes, its the same on all of my inductors. i did measure some other randome inductors i had lying around and they showed similar behaviour.
as you have said, the only proper way to fix this is most likely to re-calculate for another lower value  inductor. while a pain for all, i think it would be worth it.  if you dont have the time i might try myself but im horrible at math but building an eq that i intend to use for high end bossing  that has this behaviour is not an option for me.

If only it were as simple as recalculating for a different inductor. The problem, as I have explained privately to Jonas is one of sharpness (Q) and impedance. The sharpness (Q) of a series resonant circuit like the one used in this EQ is given by:

Q = w*L/R

where w(omega) is 2*pi*f where f is the frequency, L is the inductance and R the obverall circuit resistance. If we just make L smaller to raise the self resonant frequency, the Q is lowered (which probably explains why there are no vintage high Q EQs)

Rearranging the above formula you get:

L = Q*R/w which means for a given Q and frequency, L depends on the overall circuit impedance. The original Pultec is designed as a 600 ohm input and uses a 10K pot and a 1K pot to form its basic divider and it is these that determine the basic circuit resistance R. Today we use 10K inputs so when I designed the Poor Man's Pultec I raised the pot values by 5 times to 50K and 5K which means the circuit resistance is also 5 times higher so the inductor values need to be five times higher to get the same Q -a dn now we know what the hidden consequences of that are. The orignal Poor Man's Pultec had no inductors in it. After resisiting a lot of requests I finally relented and added  the mid section but it was always known to be a bit of a compromise. Later after many requests for a high boost like the original I came up with the 3band version. As as been said, many people have built this and are perfectly happy with it, quirks and all.

So, bottom line is, to lower the inductor value and maintain the original Q needs a complete redesign and will change the input impedance of the EQ.  The question is how low to go? The VTB9044 used in the REDD EQ I am told hs a much higher self resoance. The REDD EQ uses 22K and 5K1 resistors in its pot divider so its circuit resistance is about half that of the 3B Pultec. This combined with the inductance being about one fifth that of the Pultec means the Q of the REDD is about 2.5 times lower than the Pultec. So we could change the pot divider to 22K and 2K2 to maintain the range of the boost/cut, use the VTB9044 and accept a lower Q. To keep the Q as it is with the VTB9044 we would need to change the pot divider to 10K and 1K just like the orignal Pultec which means the input is now  more like 600 ohms and the inout trnaformer would need to be changed.

Another alternative is to keep the existing pot divider and change the inductors to use individual 100mH inductors (Toko, Murata etc). To make the 1H total of the VTB9042 will need 10 in series nd for the 2H of the VTB9050 will need 20. I think it would make sense to accept lower Q at the lower frequencies of both bands so that maximum inductance is halved in both cases so we just need just 5 inductors for the high boost and 10 for the low.

The only other thing worth thinking about is if there is some way to dampen the out of band peak (ascillation is not the right word) but doing this without affecting the the overall response may not be possibe.

Cheers

Ian

 

[salomonander]

hey Ian,

thanks for the reply. i will need some time to understand everything properly
today i went to tghe surplus shop and bought some more inductors. it was difficult to find some values that resemble that of the carnhills but i found three that had a similar tap and were of the same type of construction.
one of them behaved exactly like the carnhills. meaning that the response resembled more of a tight notch. the other two measured rather different. i attached some pics. they are not the exact same value as the carnhills but very close. maybe this helps somehow.

 

[salomonander]

160mh carnhill

 

[salomonander]

166mh siemens

 

[salomonander]

220mh carnhill

 

[salomonander]

250mh noname

 

[salomonander]

as i have said - im really no expert. but the none carnhill inductors look so much more heathy and are almost the same value. say id find a multitap inductor that has the necessary values and behaves more like the ones i have found at the surplus - would my q change? aaron at don audio does some great custom inductors for fair prices. maybe his work better.

 

[kambo]

can you measure the resistances on same value inductors.... which changes the Q