[gareth magennis]

Right, I have this Pultec EQP-1A clone that apparently was motorboating, but only when attached to a patchbay.
Turns out the manufacturer knew certain serial numbered units were prone to this and offered a mod to fix the problem.
Customer took the unit to a tech who carried out the mod as described in an email from the company.
Only now the unit is something like 12dB down in level than before.

I have removed the mod, and there is in fact about 12dB drop in level with it fitted, the customer is correct.

Original Pultec schematic he
http://www.gyraf.dk/gy_pd/pultec/pultech.gif


The mod is to wire a 100 ohm resistor in parallel with a 0.1uF capacitor directly accross the output transformer secondary.

The pultec schematic shows a 1k resistor in series with a capacitor on the output, the unit is question has a factory fitted 2k2 resistor across the output, but no cap.

I am not having much joy from the company on this matter, but it seems to me that putting 100 ohms across the output transformer is going to reduce the output level rather drastically.
This transformer is directly connected to pins 2 & 3 of the output XLR, so essentialy there is 100 ohms across pins 2 & 3.
Transformer secondary measures 113 ohms DC disconnected.

I have been trying to suggest that either this 100 ohms is incorrect, or the 100 ohms in parallel with the 0.1uF capacitor is not intended to be put across the output, but should in fact be installed somewhere else.
Maybe on the grounding circuit to prevent the motorboating under certain conditions I can't create on the bench.


Any thoughts on this? I'm getting rather frustrated.


Cheers.

[Don Pearce[_3_]]

On Tue, 2 Feb 2021 13:42:49 -0800 (PST), gareth magennis
wrote:

Right, I have this Pultec EQP-1A clone that apparently was motorboating, but only when attached to a patchbay.
Turns out the manufacturer knew certain serial numbered units were prone to this and offered a mod to fix the problem.
Customer took the unit to a tech who carried out the mod as described in an email from the company.
Only now the unit is something like 12dB down in level than before.

I have removed the mod, and there is in fact about 12dB drop in level with it fitted, the customer is correct.

Original Pultec schematic he
http://www.gyraf.dk/gy_pd/pultec/pultech.gif


The mod is to wire a 100 ohm resistor in parallel with a 0.1uF capacitor directly accross the output transformer secondary.

The pultec schematic shows a 1k resistor in series with a capacitor on the output, the unit is question has a factory fitted 2k2 resistor across the output, but no cap.

I am not having much joy from the company on this matter, but it seems to me that putting 100 ohms across the output transformer is going to reduce the output level rather drastically.
This transformer is directly connected to pins 2 & 3 of the output XLR, so essentialy there is 100 ohms across pins 2 & 3.
Transformer secondary measures 113 ohms DC disconnected.

I have been trying to suggest that either this 100 ohms is incorrect, or the 100 ohms in parallel with the 0.1uF capacitor is not intended to be put across the output, but should in fact be installed somewhere else.
Maybe on the grounding circuit to prevent the motorboating under certain conditions I can't create on the bench.


Any thoughts on this? I'm getting rather frustrated.


Cheers.


Motor boating is normally caused by taking negative feedback from the
secondary of the output transformer. It is the low frequency phase
shift as the transformer runs out of puff that causes it. But this
circuit appears to have no feedback of any kind so it is hard to see
what could motorboat. They already have a Zobel network on the output
to keep things tame at the top end.

If you are losing 12dB when you add the 100 ohms, this thing has a
seriously high output impedance (caused by the lack of feedback). I
think the approach here is to see how high a resistor value you can
get away with before motorboating starts. Then perhaps drop that by
one value and stick with that. Valves are very variable, so it is very
unlikely that every unit will need so drastic a fix as 100 ohms.

d

--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

[gareth magennis]

On Tuesday, 2 February 2021 at 22:01:06 UTC, Don Pearce wrote:
On Tue, 2 Feb 2021 13:42:49 -0800 (PST), gareth magennis
wrote:

Right, I have this Pultec EQP-1A clone that apparently was motorboating, but only when attached to a patchbay.
Turns out the manufacturer knew certain serial numbered units were prone to this and offered a mod to fix the problem.
Customer took the unit to a tech who carried out the mod as described in an email from the company.
Only now the unit is something like 12dB down in level than before.

I have removed the mod, and there is in fact about 12dB drop in level with it fitted, the customer is correct.

Original Pultec schematic he
http://www.gyraf.dk/gy_pd/pultec/pultech.gif


The mod is to wire a 100 ohm resistor in parallel with a 0.1uF capacitor directly accross the output transformer secondary.

The pultec schematic shows a 1k resistor in series with a capacitor on the output, the unit is question has a factory fitted 2k2 resistor across the output, but no cap.

I am not having much joy from the company on this matter, but it seems to me that putting 100 ohms across the output transformer is going to reduce the output level rather drastically.
This transformer is directly connected to pins 2 & 3 of the output XLR, so essentialy there is 100 ohms across pins 2 & 3.
Transformer secondary measures 113 ohms DC disconnected.

I have been trying to suggest that either this 100 ohms is incorrect, or the 100 ohms in parallel with the 0.1uF capacitor is not intended to be put across the output, but should in fact be installed somewhere else.
Maybe on the grounding circuit to prevent the motorboating under certain conditions I can't create on the bench.


Any thoughts on this? I'm getting rather frustrated.


Cheers.


Motor boating is normally caused by taking negative feedback from the
secondary of the output transformer. It is the low frequency phase
shift as the transformer runs out of puff that causes it. But this
circuit appears to have no feedback of any kind so it is hard to see
what could motorboat. They already have a Zobel network on the output
to keep things tame at the top end.

If you are losing 12dB when you add the 100 ohms, this thing has a
seriously high output impedance (caused by the lack of feedback). I
think the approach here is to see how high a resistor value you can
get away with before motorboating starts. Then perhaps drop that by
one value and stick with that. Valves are very variable, so it is very
unlikely that every unit will need so drastic a fix as 100 ohms.

d

--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus


Ok, thanks.

I haven't experienced this motorboating because apparently it only happens in certain known production units, and only when connected to a patchbay.
Presumably all those long cables and earth loops are a factor.

I doubt I will be able to make it motorboat on the bench, it appears to be fine.


Also the schematic I posted is of the original Pultec, I don't have the schematic of this clone, which is not exactly the same.
They don't have a zobel network on the output, just a 2k2 resistor across the output, no capacitor.
And obviously they use different transformers so most componet values are going to be different.


I will experiment with different resistor values to see when things start to go bad.


The problem is, the customer has a pair of these, only one of which exhibited the motorboating, and thus only one had the mod done.
Now he has one unit 12 db lower than the other.
Not good.

[[email protected]]

gareth magennis wrote:
===================

Any thoughts on this? I'm getting rather frustrated.


** Motorboating in audio stages that do not use NFB is caused by supply rail coupling of imposed signal from the later stages back to the first stage - usually at a sub sonic frequency.

The inter-stage coupling caps are way too large at 270nF - try 22nF instead for each.

This shifts the -3dB point up to to 7 Hz.

Increasing the PSU filter caps - by paralleling with a 47uF - will also help reduce LF coupling.


...... Phil

[gareth magennis]

On Tuesday, 2 February 2021 at 23:13:16 UTC, wrote:
gareth magennis wrote:
===================

Any thoughts on this? I'm getting rather frustrated.

** Motorboating in audio stages that do not use NFB is caused by supply rail coupling of imposed signal from the later stages back to the first stage - usually at a sub sonic frequency.

The inter-stage coupling caps are way too large at 270nF - try 22nF instead for each.

This shifts the -3dB point up to to 7 Hz.

Increasing the PSU filter caps - by paralleling with a 47uF - will also help reduce LF coupling.


..... Phil



Yes, customer was reporting a "tick" every couple of seconds.

[gareth magennis]

On Tuesday, 2 February 2021 at 23:15:30 UTC, gareth magennis wrote:
On Tuesday, 2 February 2021 at 23:13:16 UTC, wrote:
gareth magennis wrote:
===================

Any thoughts on this? I'm getting rather frustrated.

** Motorboating in audio stages that do not use NFB is caused by supply rail coupling of imposed signal from the later stages back to the first stage - usually at a sub sonic frequency.

The inter-stage coupling caps are way too large at 270nF - try 22nF instead for each.

This shifts the -3dB point up to to 7 Hz.

Increasing the PSU filter caps - by paralleling with a 47uF - will also help reduce LF coupling.


..... Phil
Yes, customer was reporting a "tick" every couple of seconds.


How can I force this motorboating to happen on my bench, so I can deal with it?
I've not seen it yet.

[Mike Rivers[_2_]]

On 2/2/2021 6:35 PM, gareth magennis wrote:
How can I force this motorboating to happen on my bench, so I can deal with it?
I've not seen it yet.

If it only happens when connected to a patchbay, it's likely due to the
cable capacitance hanging across the output.

You should ask the owner with the problem a couple of questions:

1. How long are the cables between the EQ and the patchbay? If it's
going to a normalled jack, add the length of the cable between the
normalled jack of the pair to wherever it goes.

2. If it's going to a normalled jack, what's that jack going to?

or

Does it motorboat if the normalled connection is broken (stick a cable
in the "destination" patchbay jack)

You could experiment with loading the output with a capacitor. Start
with 500 pF and go up in value to about 0.01 uF to see if you can make
it take off.

I agree than 100 ohms sounds like a pretty hefty load, but you can see
what the source impedance really is by putting in a tone and measuring
the open circuit voltage at the output. Then hang resistance across it
until the voltage drops in half (6 dB if you're using a meter calibrated
in dB). The resistance value that gives you the 6 dB drop will be
representative of the output impedance of the unit.



--
For a good time, call http://mikeriversaudio.wordpress.com

[[email protected]]

gareth magennis is super annoying :

=========================

How can I force this motorboating to happen on my bench, so I can deal with it?
I've not seen it yet.


** FFS !!!

How will you ever know if you have fixed it ?????

Rule one of troubleshooting: " Observe the Fault "



...... Phil

[gareth magennis]

On Wednesday, 3 February 2021 at 01:48:00 UTC, wrote:
gareth magennis is super annoying :

=========================

How can I force this motorboating to happen on my bench, so I can deal with it?
I've not seen it yet.
** FFS !!!

How will you ever know if you have fixed it ?????

Rule one of troubleshooting: " Observe the Fault "



..... Phil




Ok, I think I may have got to the bottom of this.

I found this, which I believe is the issue and its solution.
https://www.gearslutz.com/board/so-m...fications.html


There, it states the mod is a 25pf in series with 100 ohms, which makes much more sense.
I fitted it and, of course, it doesn't drop the level by 12dB.

I did measure things before finding this, and the output inpedence of the circuit is around 360 ohms, in the method described by Mike.
Thanks MIke.
I also tried to measure the low frequency rolloff, which was way below 10 Hz.
Thanks, Phil.

Maybe there never was any "motorboating", this notion came from the customer in his email to the company.
I never found any, despite adding caps on the output like Mike suggested.

[[email protected]]

gareth magennis is soooo annoying :

==============================
Any thoughts on this? I'm getting rather frustrated.

** The feeling is mutual - pal.


** Motorboating in audio stages that do not use NFB is caused by supply rail coupling of imposed signal from the later stages back to the first stage - usually at a sub sonic frequency.

Yes, customer was reporting a "tick" every couple of seconds.

** How the hell is a tick = sub sonic oscillation?

The term "motor-boating" is kinda obvious.
Got any idea what "hiss", "hum" and buzz" actually sound like ?
Know why they have those names ?


....... Phil

[Scott Dorsey]

gareth magennis wrote:
Right, I have this Pultec EQP-1A clone that apparently was motorboating, but only when attached to a patchbay.
Turns out the manufacturer knew certain serial numbered units were prone to this and offered a mod to fix the problem.
Customer took the unit to a tech who carried out the mod as described in an email from the company.
Only now the unit is something like 12dB down in level than before.

I have removed the mod, and there is in fact about 12dB drop in level with it fitted, the customer is correct.

Original Pultec schematic he
http://www.gyraf.dk/gy_pd/pultec/pultech.gif


The mod is to wire a 100 ohm resistor in parallel with a 0.1uF capacitor directly accross the output transformer secondary.

You sure they aren't in series? If they were in series, that would sound like
a normal zobel network designed to keep the transformer from ringing or to add
a stabilization pole for something else. In parallel they would just swamp
the transformer which is designed for a 600 ohm load.

I bet, given the description, that the thing works perfectly fine with a
600 ohm load, and only motorboats when connected to modern high-z
transformerless inputs. I bet the transformer sounds better with the
600 ohm load too. The Pultec design is intended for a 600 ohm load.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[gareth magennis]

On Thursday, 11 February 2021 at 12:53:54 UTC, Scott Dorsey wrote:
gareth magennis wrote:
Right, I have this Pultec EQP-1A clone that apparently was motorboating, but only when attached to a patchbay.
Turns out the manufacturer knew certain serial numbered units were prone to this and offered a mod to fix the problem.
Customer took the unit to a tech who carried out the mod as described in an email from the company.
Only now the unit is something like 12dB down in level than before.

I have removed the mod, and there is in fact about 12dB drop in level with it fitted, the customer is correct.

Original Pultec schematic he
http://www.gyraf.dk/gy_pd/pultec/pultech.gif


The mod is to wire a 100 ohm resistor in parallel with a 0.1uF capacitor directly accross the output transformer secondary.

You sure they aren't in series? If they were in series, that would sound like
a normal zobel network designed to keep the transformer from ringing or to add
a stabilization pole for something else. In parallel they would just swamp
the transformer which is designed for a 600 ohm load.

I bet, given the description, that the thing works perfectly fine with a
600 ohm load, and only motorboats when connected to modern high-z
transformerless inputs. I bet the transformer sounds better with the
600 ohm load too. The Pultec design is intended for a 600 ohm load.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."


One of the first things I did was to try them in series, in case this was the error.
This produced a significant rolloff at 10 KHz, so I guessed that wasn't the error.

The Gearslutz post says 100 Ohms in series with 25pf, which I thought made a lot more sense.
I'm reluctant to accept that without comfirmation from the company, though, which I still haven't received.

Is that a typo? should it be 22pF, or even 22nF? I struggled to find a 25pF cap.
Hence my reluctance.

[Scott Dorsey]

gareth magennis wrote:
The Gearslutz post says 100 Ohms in series with 25pf, which I thought made a lot more sense.
I'm reluctant to accept that without comfirmation from the company, though, which I still haven't received.

Is that a typo? should it be 22pF, or even 22nF? I struggled to find a 25pF cap.
Hence my reluctance.

Put a 1 KHz square wave through. Does it look nice and square at the output or
does it ring?

Mild ringing: http://www.panix.com/~kludge/xformers/DSC00263.JPG

Catastrophic ringing: http://www.panix.com/~kludge/xformers/DSC00258.JPG

If you add a 600 ohm termination resistor (which everybody should be using
running a Pultec into a modern output anyway) does it get better or worse?

If the square wave looks good, things are good.

You can use the calibration output from a Tek scope as a 1kc square wave
generator. It's convenient and comes free with the scope.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[gray_wolf]

On 25/02/2021 7:01 pm, Scott Dorsey wrote:
gareth magennis wrote:
The Gearslutz post says 100 Ohms in series with 25pf, which I thought made a lot more sense.
I'm reluctant to accept that without comfirmation from the company, though, which I still haven't received.

Is that a typo? should it be 22pF, or even 22nF? I struggled to find a 25pF cap.
Hence my reluctance.

Put a 1 KHz square wave through. Does it look nice and square at the output or
does it ring?

Mild ringing: http://www.panix.com/~kludge/xformers/DSC00263.JPG

Catastrophic ringing: http://www.panix.com/~kludge/xformers/DSC00258.JPG

If you add a 600 ohm termination resistor (which everybody should be using
running a Pultec into a modern output anyway) does it get better or worse?

If the square wave looks good, things are good.

You can use the calibration output from a Tek scope as a 1kc square wave
generator. It's convenient and comes free with the scope.
--scott


Any idea what the catastrophic ringing harmonics would look like on a spectrum
analyzer.

[Don Pearce[_3_]]

On Fri, 26 Feb 2021 02:29:37 -0600, gray_wolf
wrote:

On 25/02/2021 7:01 pm, Scott Dorsey wrote:
gareth magennis wrote:
The Gearslutz post says 100 Ohms in series with 25pf, which I thought made a lot more sense.
I'm reluctant to accept that without comfirmation from the company, though, which I still haven't received.

Is that a typo? should it be 22pF, or even 22nF? I struggled to find a 25pF cap.
Hence my reluctance.

Put a 1 KHz square wave through. Does it look nice and square at the output or
does it ring?

Mild ringing: http://www.panix.com/~kludge/xformers/DSC00263.JPG

Catastrophic ringing: http://www.panix.com/~kludge/xformers/DSC00258.JPG

If you add a 600 ohm termination resistor (which everybody should be using
running a Pultec into a modern output anyway) does it get better or worse?

If the square wave looks good, things are good.

You can use the calibration output from a Tek scope as a 1kc square wave
generator. It's convenient and comes free with the scope.
--scott


Any idea what the catastrophic ringing harmonics would look like on a spectrum
analyzer.

The same is non-catastrophic but bigger. That is just a judgement
call. The way you judge it to be actual ringing is that one harmonic
is much bigger than any other. That is the ringing frequency.

d

--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

[Scott Dorsey]

gray_wolf wrote:
kludge wrote:

Catastrophic ringing: http://www.panix.com/~kludge/xformers/DSC00258.JPG

Any idea what the catastrophic ringing harmonics would look like on a spectrum
analyzer.

They aren't harmonics!

They are the result of parasitic LC networks caused by the series inductance
of the transformer winding and the capacitance between turns. In the case
of simply-wound transformers (like the one in that picture), the result is
one big narrowband resonance somewhere. Any signal anywhere near that
resonance excites it, even the harmonics of the 1kc. And once it gets excited,
it will continue bouncing back and forth at its resonant frequency until
the energy is gone and that capacitor is discharged.

So on a spectrum analyzer that would likely be one single narrow spike.

Now... there are better ways to wind transformers in order to distribute
that across a wider area, so that you get more smaller resonances instead
of one big one. Those can result in multiple smaller spikes.

If the transformer is designed well, the spike is well above the audible
region where it can be eaten by a zobel network.

That severely ringing transformer was designed for communications applications
where having a very high ratio was more important than bandwidth or low
distortion, so that resonance is much more extreme and much lower than anyone
would tolerate for a pro audio transformer. But, this is a 200:80k ohm
transformer.... the fact that it works at all is a miracle.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."