Grounding?

for Dynaco Mark II/III/IV and DIY PP monoblocks

Grounding?

Postby Blair » Sat Nov 15, 2008 6:01 pm

This one is beyond me. I just finished my "dual mono" Poseidon. I went to ground the OPT on the common as the manual says to do. I can ground the right channel and it is fine, but if I ground the left channel, it makes an awful sqelching sound. Any ideas? It still has some hum which is beyond me because it is the same components as I used minus the PS caps for the last build and it used to be quiet. Any obvious reasons for hum or the squelching sound?

A few notes are that the IEC goes directly to the chassis and nowhere else, and the entire circuit is grounded via, RCA ground from the preamp which is a Parasound PHP-850. It is a very quiet preamp typically.

Image

Image

This is what it looks like up top with the recessed 9 pin tubes. I will probably drop the another 1/4 inch. I wanted them to look more recessed.

Image

Thanks if anyone can help,

Blair
Blair
KT88
 
Posts: 1272
Joined: Thu Nov 09, 2006 10:22 am

Postby TomMcNally » Sat Nov 15, 2008 6:08 pm

Blair -

Nice looking amp !

You have the primary (or secondary) of the output transformer
out of phase, and when the common gets grounded, the feedback
starts to "work" and it's backwards, to you get that howl.
Connect the transformer up like the other one ... reverse
either the primary or the secondary - but NOT both.

I've done that a few times - there is a disclaimer in the
front of the manuals about that. For example, the
Dyanaco/Triode transformers and the Hammonds are
backward, so it's something that may be trial and error.
User avatar
TomMcNally
Darling du Jour
 
Posts: 2729
Joined: Sat Nov 19, 2005 2:19 pm
Location: Northfield, NJ

Postby Blair » Sat Nov 15, 2008 6:17 pm

Thanks Tom. I'm sure many are tired of seeing this amp, but this is a different build. I completely stripped it and rebuilt it to clean it up. Just to make sure I follow what you are suggesting, I need to swap the blue set and brown set on the EDCOR OPT?

http://www.edcorusa.com/products/transf ... -4_2k.html

I can do this pretty easily back at the UL switch.

Thanks,

Blair
Blair
KT88
 
Posts: 1272
Joined: Thu Nov 09, 2006 10:22 am

Postby TomMcNally » Sat Nov 15, 2008 6:32 pm

Yep - you'll need to swap the UL wires also.
User avatar
TomMcNally
Darling du Jour
 
Posts: 2729
Joined: Sat Nov 19, 2005 2:19 pm
Location: Northfield, NJ

Postby Blair » Sat Nov 15, 2008 6:40 pm

OK,

That fixed that. I guess all that is left is trying to track down the hum.

Thanks Tom!

Blair
Blair
KT88
 
Posts: 1272
Joined: Thu Nov 09, 2006 10:22 am

Postby Blair » Sat Nov 15, 2008 6:44 pm

Tom,

Would it cause a ground loop if I were to just run a jumper from the RCA ground over to the common OPT speaker terminal? I would not think so considering everything grounds from the RCA. It is kind of the "Star" point right?

Blair
Blair
KT88
 
Posts: 1272
Joined: Thu Nov 09, 2006 10:22 am

Postby TomMcNally » Sat Nov 15, 2008 7:34 pm

I don't like to have the RCA inputs near anything ... or grounded
to other things ... the way you have the RCA's on either side
of the 115vac is a little too close for my taste, and then
you have the leads running across the DC supply.

I'd test this by removing the inputs from the driver boards
and shorting them and listening for hum.
User avatar
TomMcNally
Darling du Jour
 
Posts: 2729
Joined: Sat Nov 19, 2005 2:19 pm
Location: Northfield, NJ

Postby Blair » Sat Nov 15, 2008 7:50 pm

I'll give that a try tomorrow. Do the boards not pretty much automatically ground the RCAs to the ground plane?

In the mean time, do you know what causes this?

One tube on each side is like this. They all measure -58v on the bias:

Image

Thanks again for all you help Tom,

Blair
Blair
KT88
 
Posts: 1272
Joined: Thu Nov 09, 2006 10:22 am

Postby Ty_Bower » Sat Nov 15, 2008 8:16 pm

Red plate is caused by excessive dissipation. It might be due to high plate voltage, insufficient negative bias voltage, DC leaking through the coupling capacitors, or possibly because of too high a value for the grid leak resistor.

When you say each tube "measures -58 on the bias", I assume you mean you have the voltmeter probe on pin 6 of the KT88 sockets? What kind of voltage drop do you have across the 10 ohm sense resistor from each KT88's cathode to ground?
User avatar
Ty_Bower
KT88
 
Posts: 1494
Joined: Wed Mar 21, 2007 2:50 pm
Location: Newark, DE

Postby TomMcNally » Sat Nov 15, 2008 9:28 pm

Blair -

Yep - like Ty says, check the voltage across the
10 ohm resistors, it should be adjusted to .7 volts
(700 mv) at idle, with input turned down and
speakers connected. The tubes should cool down.
User avatar
TomMcNally
Darling du Jour
 
Posts: 2729
Joined: Sat Nov 19, 2005 2:19 pm
Location: Northfield, NJ

Postby Blair » Sat Nov 15, 2008 9:31 pm

I see. The plate voltage on all tubes is around 510v.

Pin 6 on the KT88 socket is non terminated isn't it?

1 & 8 tie together and ground through a 10 ohm resistor.

2 & 7 are filaments

3 is plate

4 is screen grid

5 goes to point G on the board through a 1K resistor which is where I measure the -58v. Does this make sense? I'm tube literate enough to get myself in trouble Yellow_Light_Colorz_PDT_02

Should I start by checking the voltage on either side of the .1uf caps?

Blair
Blair
KT88
 
Posts: 1272
Joined: Thu Nov 09, 2006 10:22 am

Postby Ty_Bower » Sat Nov 15, 2008 9:49 pm

Pin 5 is the control grid. It needs a 1K ohm grid stopper resistor tied close to it. The manual suggests using the adjacent pin 6 as a tie-in point; mount the stopper resistor between pins 5 and 6, and attach the wire from point F/G of the Poseidon board to the end of the resistor at pin 6. For some reason, people don't seem to want to do this and prefer to fly the resistor directly between pin 5 and the Poseidon board. So be it.

Have you correctly attached point F to one KT88 tube, and point G to the other?

The voltage on the 12AU7 side of the 0.1 uF coupling capacitor should be about 220 volts DC. The voltage on the other side (KT88 side) should be at bias voltage (about -58 VDC).

Is it possible for only one side of the push/pull pair to be oscillating? What would cause that?
User avatar
Ty_Bower
KT88
 
Posts: 1494
Joined: Wed Mar 21, 2007 2:50 pm
Location: Newark, DE

Postby Blair » Sat Nov 15, 2008 10:16 pm

Thanks Ty,

I'll check those voltages. I did not use pin 6 as a tie in because the Poseidon manual does not reference that as a common practice. I do, however have 1K 2w resistors tied to pin 5 and then to points G and F. G and F are showing -58v as adjusted by the 5K pot. I used the balance pot to get two tubes balanced where they were both at -58v. I'll check the caps and pull the tubes and check all resistor values again.

Blair
Blair
KT88
 
Posts: 1272
Joined: Thu Nov 09, 2006 10:22 am

Postby Blair » Sat Nov 15, 2008 11:20 pm

OK,

I just went in and adjusted the 10K pot while my meter was clipped across the 10 ohm resistor until I got close to .7v. One side will drop that low and is at .7v on both tubes, but the voltages at points G and F are not -58v anymore.

Here is what I get:

V1: -64v at G with .65v across 10 ohm
V2: -68v at F with .70v across 10 ohm
V3: -63v at G with .685v across 10 ohm
V4: -66v at F with .685v across 10 ohm

What can I do to get the voltage more centered in the range for the 10K pot? Right now it is maxed out as low as it will go.

However, it did cool the tubes down. Thank you guys for that! These SEDs are all I got and I do not have the $200 to replace them.

Blair
Blair
KT88
 
Posts: 1272
Joined: Thu Nov 09, 2006 10:22 am

Postby Ty_Bower » Sun Nov 16, 2008 7:24 am

In short, I think you want to replace the 4.7K resistor in the bias circuit with a 1K part. There's some info in these threads:

http://www.diytube.com/phpBB2/viewtopic.php?t=2296

http://www.diytube.com/phpBB2/viewtopic.php?t=2539

I'd also consider pairing V2 with V4, since they are a better match for each other.
User avatar
Ty_Bower
KT88
 
Posts: 1494
Joined: Wed Mar 21, 2007 2:50 pm
Location: Newark, DE

Next

Return to poseidon

Who is online

Users browsing this forum: No registered users and 8 guests

cron