Standby Switch WARNING

for the DIY ST35, the Dynakit and every other PP EL84

Standby Switch WARNING

Postby Shannon Parks » Sat Nov 27, 2004 1:35 pm

I had a customer send me one of his new Ike monoblocks after it was consistantly blowing fuses on power up. I sent him several suggestions, which he patiently implemented to no avail, so it seemed like a good amp to debug on my bench.

It ends up being related directly to the HV center-tap implemented standby switch. After placing caps across it and using IRCLs to tame things, I figured the current was under control, so I needed to look at the secondary taps more closely. Then I about blew up my DMM a few times trying to take measurements, getting 2kV DC in the secondary from either leg to ground (with no tubes and center tap lifted, ie standby off).

At the least, this can cause a eventually blown diode which then causes a blown fuse. While having some load in there, like tube filaments, may help tame the high voltage peak, I'm not willing to risk my DMM or anyone's health on it. I'm going to discourage the use of standby switches here on out.

Those with standby switches, leave them 'on' all the time. Better yet, disconnect the center tap from the switch and connect it directly to the PCB. With the center tap grounded, I have seen no problems whatsoever. The delayed turn on benefit is small, in my opinion, to the damage that can be done otherwise.

I've researched this problem and can't find anything. Using diodes in series (two UF4007's for 2kV rating) and a high voltage cap strapped across the primary/secondary or to ground might alleviate some of this problem. But I'll see if anyone has some experience in this area and wants to share. My personal opinion has probably been made, choosing to err with caution. I tested a PA774 and saw the same thing, so this is not Hammond or diytube specific.

If you have blown fuses, be sure to check your diodes first thing. With the amp off and bled, check with your DMM on the diode setting. Touch the black probe to the band (cathode) and the red to the anode. You should see a voltage drop of about 500mV. If you swap the black and red probes, the diode should appear as an open. Otherwise, you've got a blown diode. Swap them both out and disconnect your standby switch while you are at it.

Tube gurus, opinions welcome. Let's keep this a fun, safe hobby with gear that lasts a lifetime.

Shannon
User avatar
Shannon Parks
Site Admin
 
Posts: 3764
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Re: Standby Switch WARNING

Postby Uncle Ned » Tue Nov 30, 2004 2:32 am

separks wrote:I had a customer send me one of his new Ike monoblocks after it was consistantly blowing fuses on power up. I sent him several suggestions, which he patiently implemented to no avail, so it seemed like a good amp to debug on my bench.
It ends up being related directly to the HV center-tap implemented standby switch.
Shannon



Two suggestions:

1. Put the standby switch in the DC B+ lead, and use a Carling toggle switch. Worked for
years for Fender, though they're a frequent replacement item.

2. Use a DPDT switch and put it on the AC side of the diodes (a la Marshall). The voltage
rating of the switch can be considerably lower than what's needed, the
ratings are a combination of things, generally the lower the current (up to a point), the
higher you can push the voltage, and switches dislike DC. My pop, who worked for GE
General Purpose Control in Bloomington for years, once gave me a long lecture
on switch & relay ratings, and it's more complicated than most people think.
The basic problem is that DC tends to arc over & pit switch contacts, or even
glue them together (a problem which anyone who ever had to fix a point type
ignition system can relate to), or arc over & ruin the insulation.

There are special switches & relays that are made for switching HV DC,
but unless you work for a place that fixes transmitters they're tough to find.

Any way, the best suggestion I can make is use the heaviest thing you can
find, use it on the AC side, and derate it by about 90%.
Ned Carlson
www.tubezone.net
SW Side of Chicago,IL USA
User avatar
Uncle Ned
 
Posts: 127
Joined: Tue Apr 15, 2003 8:40 pm
Location: Soggy Wastelands of Old Hampshire (for a while)

Re: Standby Switch WARNING

Postby Shannon Parks » Wed Dec 01, 2004 5:17 pm

Uncle Ned wrote:Two suggestions:

1. Put the standby switch in the DC B+ lead, and use a Carling toggle switch. Worked for years for Fender, though they're a frequent replacement item.


Thanks a ton, Ned! Is there a particular Fender model/circuit that I can Google for? Many thanks.

Shannon
User avatar
Shannon Parks
Site Admin
 
Posts: 3764
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Re: Standby Switch WARNING

Postby Uncle Ned » Wed Dec 08, 2004 3:28 am

separks wrote:Thanks a ton, Ned! Is there a particular Fender model/circuit that I can Google for? Many thanks.Shannon


D'oh! Like nearly ALL of them. The Carling SPST standby switch, like Delco V8 distributor
caps on Chevys, was standard issue on Fender amps for decades. You want some,
(switches, not distributor caps) I've got a bucketful I'll sell you for 3 bucks a pop.
They're about as reliable as any cheap switch can be for switching DC, but like
distributor points & caps, they crap out after enough use.

You want Fender schematics? Try http://www.deltabluesman.com/
Scroll down and click on Schematic Heaven, you'll find more Fender diagrams than
you can shake a slide rule (anyone remember those?) at.
Ned Carlson
www.tubezone.net
SW Side of Chicago,IL USA
User avatar
Uncle Ned
 
Posts: 127
Joined: Tue Apr 15, 2003 8:40 pm
Location: Soggy Wastelands of Old Hampshire (for a while)

Postby Alex » Wed Oct 12, 2005 5:07 am

What would happen if you were to tie a 1 Meg resistor across the standby switch? That would provide the ground reference and the startup voltage on the B+ would be pretty low as the result of the drop to merit a standby switch?
Alex
 
Posts: 13
Joined: Fri Sep 23, 2005 9:12 pm

Postby erichayes » Wed Oct 12, 2005 1:10 pm

Hi All,

Back in the days when silicon diodes were almost a luxury, manufacturers would use a device that had as little headroom as they could get away with. Since voltage spikes, rather than excessive current draw, was the usual reason a diode shorted, Fender used 2 or 3 diodes in series on each leg of the rectifier assembly. According to my former partner, who worked for Fender from 1963~1974, they rarely had problems with diode failures, although, as Ned pointed out, standby switches were popular spare parts.

I've always felt the 4000 series of diodes was barely adequate for power amp use. I've used 1N (or UF) 5408s in series of 2 or 3 per leg, and have had 0% failure in all the amps I've built--even the ones with switched B+.

Oh, and Ned... I have a bamboo 10" K&E Log Log Duplex Decitrig sitting at my side as I compose this.
Eric in the Jefferson State
erichayes
KT88
 
Posts: 987
Joined: Fri Jan 23, 2004 9:01 pm
Location: McKinleyville CA

And I have...

Postby Shannon Parks » Wed Oct 12, 2005 1:43 pm

erichayes wrote:I have a bamboo 10" K&E Log Log Duplex Decitrig sitting at my side as I compose this.


I have a Dietzgen 1757 from my dad. Back in the old days, engineers stuck out pretty easily with those babies poking out of their pockets! Yellow_Light_Colorz_PDT_11

Alex, go ahead and give it a shot. I think that may do the trick. Certainly beefed up rectifiers are a good idea, too. The bottom line is that the secondary always needs to be referenced to ground.

Shannon
Last edited by Shannon Parks on Thu Oct 13, 2005 6:22 am, edited 1 time in total.
User avatar
Shannon Parks
Site Admin
 
Posts: 3764
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Postby Uncle Ned » Thu Oct 13, 2005 2:59 am

Alex wrote:What would happen if you were to tie a 1 Meg resistor across the standby switch? That would provide the ground reference and the startup voltage on the B+ would be pretty low as the result of the drop to merit a standby switch?


What this does is allow the capacitors to charge up a bit (if you wait a bit before you flip the switch), thus reducing the voltage differential and current draw across the switch contacts. You could parallel the resistor with a small disc cap to minimize contact arcing.

Re the slide rules, I never learned to use one (and I never learned calculus, either, and thus have been a disappointment to my parents). What's amazing is that some of the best (and most reliable) amplifiers ever made were designed by guys who had to use good old common sense, pounded out calculations on slide rules and verified everything by bench tests.

HTH they manage to make cell phones work.. OTOH, I do know. There's a lot of money to be made (and lost) in cell phones, so no one just assumes the Spice models & software work right off the board.
Ned Carlson
www.tubezone.net
SW Side of Chicago,IL USA
User avatar
Uncle Ned
 
Posts: 127
Joined: Tue Apr 15, 2003 8:40 pm
Location: Soggy Wastelands of Old Hampshire (for a while)

alternative method

Postby dhuebert » Thu Oct 13, 2005 12:10 pm

Cathode Switching
The most common "new" stand-by method is called "cathode switching". This is simply a ground-interrupt switch in series with the cathode circuit of the output stage. To reduce the switch transient, and more importantly, to avoid the risk of cathode stripping, a resistance is wired across the switch contact. Any value from 4K7-2W to 10K-2W is suitable. This is equally applicable to both cathode bias and fixed bias amplifiers, with any type of output tube.



O'Connor, K. 1996 Principles of Power.Power Press Publishing. p3-25.

Great book, highly recommend!

Don
User avatar
dhuebert
KT88
 
Posts: 820
Joined: Thu May 01, 2003 9:26 am
Location: Winnipeg Manitoba Canada

Postby Alex » Tue Oct 18, 2005 6:27 pm

I tried wiring the 1M resistor across the standby switch today. So far, so good. I do not have a scope so I wouldn't see any really fast transients, but what I do see - is that the voltage at B+ rises from 0 to 6V in about 15 seconds after power switch is turned on and stays there. Each leg of the transformer secodary measured to ground is about 340 VAC - the ground reference is there.

Unless any you want to put a scope across and see if there anything is happening, but so far it seems to me the standby swich is OK to use after 1M resistor is wired across it.
Alex
 
Posts: 13
Joined: Fri Sep 23, 2005 9:12 pm

Postby sorenj07 » Mon Jun 18, 2007 12:00 am

I blew one of the fuses in my EL34 Williamson like this. The B+ comes from a 1000VCT transformer. I cut out both UF4007's and just used two in series per leg. No problems since. And yes, I do use a Carling toggle switch to connect the center-tap to ground.
User avatar
sorenj07
KT88
 
Posts: 925
Joined: Tue May 02, 2006 2:39 pm
Location: Berlin

Postby Shannon Parks » Mon Jun 18, 2007 8:30 pm

I think the 1M resistor strapped across the standby switch is the key. Ends up providing the ground reference to make sure crazy 1kV+ spikes don't happen.
User avatar
Shannon Parks
Site Admin
 
Posts: 3764
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Postby erichayes » Mon Jun 18, 2007 10:46 pm

Regarding O'Connor's cathode switching scheme . . .

This might be okay for low powered amps using 6V6s or EL84s, where the output tube plate voltage is close to the rest of the voltages in the PI and FA. But when you get into the medium-to-high power amps that are running 450~600 volts on the output tube plates, well, you're gonna have a problem with downstream caps that are only rated at 250 or 350 volts.

Since the lion's share of the power supply's capacity is consumed by the output tubes, switching them off--and leaving th PI and whatever is ahead of it connected--is going to throw (to quote a 1974 Toshiba microwave oven warning sticker) "many hundreds high volts" across said circuitry. This, to quote another (tarnished) luminary, is not a good thing.

The one meg resistor might be a trifle high, but it's certainly a good starting point. The idea is to get the resistor to work in conjunction with the IRCL so ever'thang gets heated up the way grandma used to bake her pies--nice and even.
Eric in the Jefferson State
erichayes
KT88
 
Posts: 987
Joined: Fri Jan 23, 2004 9:01 pm
Location: McKinleyville CA

Postby mesherm » Tue Jun 19, 2007 1:31 am

Eric has a point regarding cathode switching. If you take the tube load off of a Dynaco MkIII for instance, the power supply voltage will exceed the filter caps voltage rating and bad things happen.
User avatar
mesherm
KT88
 
Posts: 1232
Joined: Fri Aug 27, 2004 10:33 pm
Location: Alvin Texas

Postby Ty_Bower » Thu Jul 05, 2007 10:03 am

separks wrote:I think the 1M resistor strapped across the standby switch is the key. Ends up providing the ground reference to make sure crazy 1kV+ spikes don't happen.


Is the jury still out on this one? I'd like to use the standby switch, and I'm sure I can scrounge up a 1M resistor (or something close).
User avatar
Ty_Bower
KT88
 
Posts: 1494
Joined: Wed Mar 21, 2007 2:50 pm
Location: Newark, DE

Next

Return to stereo 35

Who is online

Users browsing this forum: No registered users and 10 guests