The official G9 help thread

[adrianbaugh]

Hi everyone, I'm after a bit of help.  I've built a G9 but it has a problem with rectification noise (a small amount of hum, but mostly rectification noise - I plugged the output into my computer and ran a frequency analysis and there was a small peak at 50Hz, a large peak at 100Hz and then harmonics of those frequencies tailing off).  The noise is really loud - generally louder than or as loud as the input signal except on the highest gain.  Changing the gain setting makes no difference to the noise, but changing the volume knob does, so I figure for some reason it mainly affects the SRPP stage.  It's exactly the same on both channels.

I tried shielding the AC supplies to the board - no success; the only unshielded cables are the ones running to the 10-way headers, for which I've used ribbon cable.  Moving the main power transformers around makes no difference.

I wondered whether it might have been coming from interference from the 240VAC wires to the terminal block running close to some components, so I tried using the extra PCB to make an offboard power supply, just feeding the resulting 245VDC to the AC input to the HV power supply on the board.  Now this dropped the onboard HV by about 20V, but the thing is - it worked.  The hum and rectification noise disappeared and the unit sounded amazing.  At least for a short while - I wanted to get the voltage back to design intent and fitted a link across the TL783 protection diode so as to bypass the onboard rectification and avoid the voltage drop.  After this (and even when I removed the link again) the noise was back, just the same as ever.

I've tried replacing the TL783s (both onboard and on the offboard power supply) and then tried the G9 with and without the offboard power supply, but with no success - it's still noisy.

So I'm pretty stumped - I'll be very grateful if anyone can offer help.

 

[MeToo2]

I have had success with an off board supply (see earlier posts in the thread)

I would not advise having both on and off-board PSU active in series. One or the other. Otherwise they'll very likely interact.

Removing R37, and disconnecting the AC in to D17/D18 would give you a point to inject your off-board supply temporarily without having to desolder too much. (C15 also appears on the same 3 pin header, so you can inject your off-board supply on this connector).

Be careful what you're doing here. And discharge the caps in advance.

What input voltage do you have to your off board supply?

If it's too low, the TL783 won't regulate, and you will get lots of mains ripple coming through to the SRPP stage.
That would match the symptoms you describe, because the first stage has it's own additional smoothing (R31/C25).

The dropout voltage of a TL783 is about 20V, so you need around 245+20 = 265 V at the input.

A temporary test could be to change the ratio of R34 to (R35+R36) to lower the supply to the anodes of the SRPP to say 220V rather than 245.
VO ≈ 1.25* (1 + R34/(R35+R36)) e.g. change the 100K R34 to 91K. or increase 47R R36 to 100R.

Before you do this make sure you're not going to exceed the max Vo - Vi on the TL783 = 125V max.

 

[adrianbaugh]

Thanks for the suggestion MeeToo2.  Input voltage measured at 269V, which should be okay.  I tried lowering the SRPP B+ anyway, by tacking a 1M in parallel with the 100K to give approx. the 91K you calculated.  The B+ reduced to 223V but the noise was essentially the same.  The B+ appears well regulated - it's a steady 245V and if I measure it with the meter in AC mode there appears to be only 0.1mV of AC.

If the C25/R31 filter is working so well on my input stage, could it indicate a problem with C15 failing to filter the DC properly?  I did have all 3 wires to the offboard caps twisted together for neatness.  Just in case that was allowing noise to propagate between the positive leads of C14/C15 I untwisted them - no difference.

 

[MeToo2]

Thanks for the suggestion MeeToo2.  Input voltage measured at 269V, which should be okay.  I tried lowering the SRPP B+ anyway, by tacking a 1M in parallel with the 100K to give approx. the 91K you calculated.  The B+ reduced to 223V but the noise was essentially the same.  The B+ appears well regulated - it's a steady 245V and if I measure it with the meter in AC mode there appears to be only 0.1mV of AC.

If the C25/R31 filter is working so well on my input stage, could it indicate a problem with C15 failing to filter the DC properly?  I did have all 3 wires to the offboard caps twisted together for neatness.  Just in case that was allowing noise to propagate between the positive leads of C14/C15 I untwisted them - no difference.

Smart move placing the 1M in parallel to the 100K for the test.

If the B+ looks stable (although 0.1mV ripple is quite a lot assuming you can measure it reliably) then you need to be looking earlier in the chain e.g. noise getting in at the output level pot.

That's a high impedance input to the SRPP grids. Twisting and routing the wires correctly here did help on my build.
If you're using shielded cables, check you haven't indadvertedly created a ground-loop that's introducing hum.

 

[adrianbaugh]

I've replaced the 10-way ribbon cable on one channel with some 9-connector shielded cable - no difference to the noise.  I noticed that if I pull the "outamp" pin on the 10-way connector on the main board to ground so as to provide a zero input to the SRPP stage - it got rid of all the noise!  So the SRPP stage actually looks okay.

I've started checking for earth loops: if I disconnect the earth line from the input XLR to the chassis I get >10M between the circuitboard earth and the chassis, so it looks like the 78S12 is properly isolated from the chassis.  Any other common earth loop problems people know of?

 

[Voodoobeat]

hi,

do these work for c14/15?
http://www.reichelt.de/Becher-Elkos/BE-100-385/3/index.html?&ACTION=3&LA=2&ARTICLE=5384&GROUPID=3146&artnr=BE+100%2F385

What confuses me, is that they actually have 3 pins (like on the second picture).  What is that 3rd one for?

Many thanks

 

[gyraf]

Will work just fine

3rd pin is just for holding it well to the pcb. Usually the writing on the capacitor describes which are which.

Jakob E.

 

[adrianbaugh]

I've no more time to work on my G9 until Friday or Saturday now, but here's a question for the next project: the 2U enclosure I bought from Farnell was a horrible flat pack thing made of really thin flimsy metal, with vents in all sides.  It does the job, but feels cheap and nasty.  If it hadn't been so expensive (for what it is!) I'd be tempted to replace it, but at least I want to find a better supplier for next time.  What do others use?

 

[bruce0]

I used a Mid Atlantic case.  Built like a brick shithouse, but it doesn't come apart easily and the screws are self tapping with no inserts so repeated opening and closing of the case causes them to strip.

I am always amazed at how poorly cases are documented on the sites that make them.  I think small run guys like us just are too much trouble.

But I am interested in the answer to this too:

What I want is a case that is:

1) Configurable (depth, 1U 2U 3U (and it would be cool to have optional half rack widths in a tray))
2) Configurable divisions within the case (so I can separate the PSU from the audio).
3) Punched with some AVAILABLE IEC connector (be nice if it was filtered fused and switched and AVAILABLE).
4) VENTED sides are nice, top and bottom vents should be at the back of a deep case (both of these allow racking of tube gear without leaving so much space)

But I couldn't find this

Lately I have been looking on ebay for rack mount telecom and pro audio equipment sorting by price.  Sometimes there are cases configured pretty well and punched and all you would need would be a front panel (throw away the guts, use the IEC entry or dc entry). Seems like prices with shipping are usually at least $30 and you have to still find a front panel. Save the landfills at the same time.

 

[Deepdark]

Hi folks

I got some questions before populate the board. I read the entire thread so I just want to be sure and free my mind of all interrogations haha. First thing first, I plan to maybe use the external spare psu board.

1. Concerning the BR1 and BR2 on the psu board, if I use a Vishay #W04G-E4/51, is it ok? It is rated 1.5A, 400V. I guess I'm ok with that.

2. Concerning the connection of the psu board, I plan to populate it entirely, not just the HT. To Cap1 + is the + lead of C14, To Cap1 - is the - lead of C14. To Cap2 + is the + lead of the C15, but 245V out, is it supposed to be the - lead of C15? Then, I run a wire from Cap2 + to the left pins of the 3 pins header of the main board and the 245V out to the center pin of the 3 pins header? Am i allright? 220VAC is my T2 primary connection and 15VAC is the secondary of T1. But, Heater out, where is it supposed to go on the main board? On the schematic it is named FF. When I look at the main board, there is a connectioon named FF 0V and +12V (just beside C105/C103). It was supposed to be 2 leads but if I connect my heater out to thOSE point, do I still have to put the lead?

3. Concerning the grounding scheme, I see a lot of different method, so I guess there is more than only one way of doing it. The way I wanna go is: Connect all XLR's pin 1 together, star ground them and connect the pin1 of the first input XLR to the GND of the main board (right pin of K1). This way, I'm sure there is no ground loop since the board is gorunded via the same ground star as the xlr and the IEC.

4. If I put a metal wall between the PSU and the main board (seen on some build) and it's bolted on the case, then I don't need to ground it since it is part of the case, Am i right?

5. Shielded wire: I use Shielded 2 conductor wire for the xlr. For all of them, exept Input 1, which is connected to K1, I don't need the ground lead because all pins 1 are tied together. So, can I just cut the ground lead on both side of the shielded wire of do I have to connect 1 side? Sorry for the newbie questions  :

See the attached pdf related to questions 1 to 3. Thanks in advance.

Regards

 

[adrianbaugh]

Fixed my noise issue by grounding the ground pins on the output XLRs. At some point I'll add a ground loop block circuit but with just a mic plugged in it's already sounding really good! 8)

 

[adrianbaugh]

1. Should be okay though I prefer to have a little more margin (remember 240V is RMS so the peak voltages will be 330V approx.  I can't remember the make of the BRs I used (could have been Vishay) but they did a series going right up to 1000V and even the 1000V ones were only 48p so I went for those.

2. I ran Cap1+ and - to one big electrolytic (mine were 400V 330uF) and + and - first to the second big electrolytic and then straight on from that to the C15 cap2+ and 0V screw terminals on the main board.  The third screw terminal of that set is then unused and you don't need to populate any of the HV section on the main board.  (I initially looked at using the 12V part of the PSU board but found that some of the tracks ran very close together and it was difficult to avoid a short in one place, so in the end I didn't bother and just stuck with the mainboard heater supply circuitry.

3. Works for me.  I know the construction guide says only to ground at the input XLRs but I had horrendous hum until I grounded the output XLR ground pins to the chassis.

4. Should be fine unless you have an enclosure covered in insulating powdercoat and a holde big enough so the bolt doesn't necessarily touch the sides.  I'd be tempted to ground it anyway.

5. If you cut the ground lead on both sides of the shielded wire the shield will be floating (and not really doing its job).  You should certainly not ground it at both ends, in order to avoid ground loops, but it ought to be grounded at one end.  I added quite a bit of shielding to some of the AC cabling away from the main board when I was trying to solve my hum problem, and for these I brought all the shields together and strapped them to ground at the negative terminal of the first big capacitor on the HV PSU, you could also just bolt them straight to the chassis at your star ground point.

By the way, I think you definitely need to connect those jumpers you queried in red.  I didn't have any of that nice insulated jumper wire (and besides, things like that make good test points) so I just connected them using cut off component legs.

 

[Deepdark]

Thanks for the reply. So, for the psu board, the way I plug my caps is the good one? If I want to keep just the HT on the psu board, according to the shematic, here is my understanding: I plug the 220 from the T2 on the 220VAC header on the psu board, I keep juste one bridge rectifier, the one replacing D17-D20, then R33-R37, D2-D5, IC1, C14 and C15 connections. Is that allright? For the C15 connection, If I run from the psu board header named "To Cap2" to the C15 pins , then run wires from C15 to the 3 pins header on the main board, leaving right pin empty. I guess it would work the same way.

 

[adrianbaugh]

Yes.  The PSU board is pretty much divided in half so if you populate the 220V screw terminal block, BR1, R33 and everything from there up to the TL783 then you will have just the HT side.

 

[adrianbaugh]

Just to say that now my G9 is working correctly I've had the chance to try it with a decent microphone (and a decent singer!!) and also with a fantastic limited edition Wal bass guitar.  It works brilliantly with both, and my wife will love it for her vocals, but as a bassist I was blown away by the gorgeous squelchy sound I could get from the Wal at high gain settings.  Thanks Jakob, not only have you provided an awesome DIY project but you've inspired me to start work on a few designs of my own.

 

[glover]

Hi, I read all 74 pages of this thread, I have use some tips and clues about hum and grounding. I have done my ground star so now I don't have hum or oscillations on output. My problem is noise in background. Preamp works as it should but even with no gain and output level on 0 I have noise on mixing desk input. I can't measure the level of noise but it's well audible.
I have correct voltages (12V, HT,Phantom). I made DI mod, double check components values on main and switching board, I made PSU on additional spare pcb (12V and HT), which is closed with power transformers and filtering capacitors in metal screening can. Don't have any ideas what else can I do with this issue. Maybe someone here had same problem and can give me some advise what can I check ? Any help is welcome !

 

[MeToo2]

Hi, I read all 74 pages of this thread, I have use some tips and clues about hum and grounding. I have done my ground star so now I don't have hum or oscillations on output. My problem is noise in background. Preamp works as it should but even with no gain and output level on 0 I have noise on mixing desk input. I can't measure the level of noise but it's well audible.
I have correct voltages (12V, HT,Phantom). I made DI mod, double check components values on main and switching board, I made PSU on additional spare pcb (12V and HT), which is closed with power transformers and filtering capacitors in metal screening can. Don't have any ideas what else can I do with this issue. Maybe someone here had same problem and can give me some advise what can I check ? Any help is welcome !

Have you properly terminated the input of the G9 e.g. with a 200 ohm source, before measuring?
Is the noise equal on both channels?
Does the noise change with different gain settings?
Can you measure the level of the noise?
Can you measure the characteristics of the noise (frequency/ spectrum)?
What mixing desk are you using?
Are you using a line in input, or a mic in input?

 

[glover]

Thanks for reply MeToo2. I use 200 Ohm resistor or dynamic microphone to terminate input. Noise is same on both channels and increase only when I put output to max, let's say from 3 to 5 o'clock of output knob. I will try to measure it and take characteristic screenshot. Preamp connected to small Phonic mixing desk and Yamaha 01v96 using line in inputs.

 

[gemini86]

I've noticed that V2 usually isn't the problem when it comes to noise (I'm assuming it's white noise we're talking about?). Changing out V1 can help. Luckily I have a small box full of 12au7s to choose from and select quiet tubes.

 

[glover]

Thanks for tip gemini86, I should get new tube's on Mondey so there will be possibility to check this thing. Below is screenshot of noise and here https://soundcloud.com/glover6/g9-noise is recording of noise from Gyraf G9, at the beginning output knob on max then goes down to zero. Input set to max but noise is same on every input setting. Don't look at level because I don't have calibrated sound card.