G-PULTEC problems, please help!

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
i've searched the forum, but what helped was playing around with grouding.

my sollution:  Connected the ground from each input together, and then to IEC ground, and give the toroid a turn and the schield around.
now it's just a slight 100 hz humm but almost not hearable with my monitors at unity gain.

people discharge the caps before you do anything... it hurts ALOT ;)
 
Hello again,

back to the question of strapping 600 Ohm across the G-Pultec output....

I've found sometimes I want to put the G-Pultec in front of either 1176 (for whatever reason).  The 1176 has 600 Ohm front end and does not play nice with the G-Pultec terminated to the front end ie: hum poor tone/level etc...

If I am to strap 600 Ohms across the output, do I lose anything sonically when I use the G-Pultec on other devices that have or want to see a higher impedance load?

Should I just make up a cable/adapter to use when I want to put the G-Pultec in front of the 1176 instead of making a permanent mod with the resistor across the output?

I think I just answered my own question and will try to rig up a simple cable that will have the 600 Ohm resistor.  Is there a downside to doing it this way?

Thanks everyone for your always helpful support!!  8)
 
Hey all  bout to build a couple of units of these, just wondering how essential the input tx is? Apart from it facilitating the balanced input are there any other benefits, does it impart any character? I thought given the input an output tx are the same that maybe I can do away with the input tx? Anyone compared, what's the verdict ?
 
No real problem in building it without input transformer - most (though not all) of sonic signature is from output transformer.

But for balanced input operation, you'll need it.

Jakob E.
 
Hey everyone. I've been desperately trying to get this baby up and running on my own and have read these 14 pages over and over but finally must admit defeat. I'm really close so hopefully it's something minor.

Right, so far I have all the correct voltages, 6.3v at heater, 240ish v into tube also, 120v into output traffo, all good. The unit in bypass works fine, as clean as if it weren't in the signal chain. When bypass is switched off, this is what im getting:

Low Cut - OK
Low Frequency select - OK
Low Boost - Nothing (always seems bass heavy though, so maybe shorted to full setting somehow?)
High frequency select - Nothing
High Boost - Nothing
Bandwidth - Nothing
High Cut - OK
High Cut Frequency - Seems to make a difference between bottom 2 settings only.

So to summarise, I think my low frequency filter pcb is working correctly with the exception of the low boost seeming to be stuck on full setting. The unit is always bass heavy regardless of where any of the dials are.
My high frequency board appears to not be doing anything apart from the high cut and MAYBE the first 2 settings on the high cut frequency selection switch.

I can't for the life of me work out whats wrong with it. My PCB link wires are correct, I've tested the pots to ensure the resistance changes with operation and they do. I've been trying to trace the path of the audio to track down the fault but am getting nowhere. From what I gather, the audio enters the filter PCB's via the IN connection below the bypass switch, and returns via the OUT connection. Literally one path in and out, but i am confused as to how the signal gets past the first pot even! Obviously it does, but to try and understand the circuit more could someone please explain the following for me. The signal enters at IN, it then travels straight through the LF PCB to the HF PCB, without going to anything. It then goes into the high boost pot, on the left hand connection. Why does that pot then not affect the level going to everything else, and if that high boost pot was turned all the way clockwise, how does any signal pass at all?

Sorry for being a noob y'all. Great site btw.
 
So I just finished my first pair of g-pultec build in a 3RU.

I did them using Haufe (turns ratio 1:1) and Pikatron (turn ratio 2:1) traffos and I used a single PSU traffo with 250VAC (no CT) and 9VAC windings (no CT) and also I used input/output relay bypass modules at the XLRs.

So, each unit was fine and I tested all good individually.  :)

When I wired them both into the psu, the noise floor gets really bad.  :(

ie. 2x g-pultec blocks wired in parallel to psu hv and heater supplies

Checked the noise floor and it is shot to hell with both modules powered on together.

Further investigation revealed the problem to be noise on the heater grounds.

After trying a lot of things (basically re-inforcing the ground trace on each g-pultec with thick
grounding wire), I found the major fix :

    - connecting the "-ve" pin of the heater rectifier bridge of each channel direct to chassis star ground

Noise floor ended up at -83.5dBu RMS (averaged over 20-20KHz) improved from -60dBu before mods. Thats with eq engaged and flat (loop back of Motu interface gives -84.5dBu)

So now, I have dual g-pultec modules in a 3RU rack with a personal best noise floor.

Frequency response is flat as a pancake at  < +/- 1dBU at 20Hz and 20KHz
Curves are all as expected.
Overall gain (eq flat) of my implementation is around -3.6dBu with ECC88/6922

Sound is very pure - not at all 'tubey'. Going to sub the 6922 (which I don;t have many)
for russian 6n5p (which I have lots) a have a listen shortly.

Maybe a few more tweaks to come.  :D :D :D


Cheers
 

Attachments

  • pultec1.JPG
    pultec1.JPG
    61.3 KB · Views: 114
Hi Dan

About your problem - when you have switch positions which seem to not work, it's usually bad solders
or hairline fractures of the trace at the pad. These single sided pcbs are not the thickest traces
and the tracks can feel some stresses when you push the control pcb thru the holes of your front panel

(if they don't line up correctly and are at the right height off the pcb)

Check carefully the solders at the switches and pots as well as at the caps and then inductors.

Use the continuity test of your multimeter to make sure the connections make it from
on pad to the next.

Make sure your connections between the 2 halves of the control pcb are properly connected
and not intermittent or loose.

You can  narrow down the problem by identifying exactly which positions don't work
and correlating with the circuit diagram component.

I also once had an inductor open circuit which caused a couple of positions not to work.

By the way, how do you test? Some of the eqs can be subtle even at full boost/cut, depending on
the souce material.

Best way is to use a PC sound card with spectrum analyzer application which can do frequency sweeps.
That way you can exactly check each position and each cut/boost/Q knob.

A simple audio probe connected to pc powered speakers is also very helpful in tracking the audio path
(careful not to put it on high voltage traces though !!) so as to compare with schematic.

Cheers
 
Hey guys,

Anybody have a recommendation for a GPultec output transformer that would be good for interfacing with 600ohm loads as well has higher impedance loads?

Is the Edcor XS1100 versatile in this regard? Seems like I read on some thread that there were issues when a 600 ohm load was connected to the GPultec out. The Edcor data says it's good for >600ohm loads, so . . . . .
 
I've used Carnhills as output transformers on Gyrafs srpp stages
as well as a number of other tube line amp stages.

I like them very much in tube outputs. The price is very reasonable from Colin at Audio Maintenance.
Similar to the Edcor XSM - maybe a little larger.

For 2:1 the VTB2291 ungapped is 2400/9600:150/600 so 2400:600 works.

They also work out well for 4:1 9600:600 applications

Apart from that, I've tried the Lundahl which is very transparent but very expensive,
the Edcor which works quite well and is cheap
as well as Haufe/Pikatron reused transformers - very transparent and can be cheap.

If I were going again, I'd pick some UTC. Sometimes you can find a suitable 1W max level 2:1
for a price at or below lundahls.

 
Dual G-pultecs seems to be notoriously problematic regarding hum - we don't recommend doubling up like that.

Have you tried looking through previous posts here?

Jakob E.
 
Grounding is an art in itself ,  if not star type than separate signal from psu grounds
but have the psu grounds meet at a chasis point near the power xfmr , especially any center taps ,
Study grounding , get tad to help you . good luck
 
yes, if spend my sunday reading over here offcourse ;)
but i don't know what i can try or not before i dammage it

i've spend hours with grounding, now only inputs go to pcb ground and iec to chasis.
that gives me almost no 50hz.

I've seen a picture somewhere from a dual unit witch shares 1 6.3v psu section for both the units.

i guess i can try that but use a bigger heatsink, cause it will double current demand on the lm317 reg..?

 
Use LM350 instead of LM317, and higher current diodes instead of the tiny rectifier bridge (fast or schottkey are best). If you are rectifying 12VAC and then regulating down to 6.3V, you can modify the PSU and include a CRC filter prior to the regulator(instead of the single 4700uF cap).  No guarantees that this will solve your problem, but it did help me with some SB4000's (with stock PSUs they hummed, until I upped the rectifier and pre-reg caps, but this is brasil, our power sucks down here). 

2 ECC88's in series will use 730mA, a HUGE load for an LM317 which maxes at 1.5A (ideally with sufficient heatsinking) and 1A diodes. Give your heaters more breathing room.  That's also the case with the power trafo, make sure it's sufficiently spec'ed.

Hope that helps, if not, it'll only have cost you 5-10 euros.
 
Hi schrobbelbop,

      Not sure what else I can add to what Jakob, okgb and mitsos have already stated. I was able to build a Dual unit , ( see page 6 of this thread )  and had planed on splitting them up in to separate units if I ended up with a Hum/Buzz problem , but I lucked out and never did .

  I used a single PT that had  a single HT secondary  250v ct @ 130mA and a single 6.3v  heater secondary @2A  . 

    I split both secondaries from a terminal strip to each SRPP  PCB where I used a 33k 5w  dropping resister for the the HT so it sat at 245vDC  , and a TL1086 in place of the lm317 ( you'll have to check the specs , not sure you will be able to use this one )  the only reason I used it was the heater seconday was already 6.3vac - so rectified it's like  8.8vDC  - was not enough " differential"  to make the lm317 work  -  also  Kingston mentioned that just a dropping resistor could be used in this situation too.

  .. so anyway ,  If you look at the pics in the link in my post on page 6 ,  my XLR pin 1's are shild grounds from the SRPP PCB's  only. IEC and PT ground to Star ground stud only and thats it .

    I have another ground wire attached to Start ground ... I was going to connect to XLR pin 1's , however I've never hooked it up because so far I haven't had any grounding problems.


  I guess its just NewBee Luck ...  Hope I was able to help

Please let us know of your progress
 
alexc said:
So I just finished my first pair of g-pultec build in a 3RU.

I did them using Haufe (turns ratio 1:1) and Pikatron (turn ratio 2:1) traffos and I used a single PSU traffo with 250VAC (no CT) and 9VAC windings (no CT) and also I used input/output relay bypass modules at the XLRs.

So, each unit was fine and I tested all good individually.  :)

When I wired them both into the psu, the noise floor gets really bad.  :(

ie. 2x g-pultec blocks wired in parallel to psu hv and heater supplies

Checked the noise floor and it is shot to hell with both modules powered on together.

Further investigation revealed the problem to be noise on the heater grounds.

After trying a lot of things (basically re-inforcing the ground trace on each g-pultec with thick
grounding wire), I found the major fix :

    - connecting the "-ve" pin of the heater rectifier bridge of each channel direct to star ground.

Noise floor ended up at -83.5dBu RMS (averaged over 20-20KHz) improved from -60dBu before mods. Thats with eq engaged and flat (loop back of Motu interface gives -84.5dBu)

So now, I have dual g-pultec modules in a 3RU rack with a personal best noise floor.

Frequency response is flat as a pancake at  < +/- 1dBU at 20Hz and 20KHz
Curves are all as expected.
Overall gain (eq flat) of my implementation is around -3.6dBu with ECC88/6922

Sound is very pure - not at all 'tubey'. Going to sub the 6922 (which I don;t have many)
for russian 6n5p (which I have lots) a have a listen shortly.

Maybe a few more tweaks to come.  :D :D :D


Cheers

  ... just saw this post , BEAUTIFUL work alexc !!  nice piece of gear  8)
 

Latest posts

Back
Top