Soft Start or JAS (Just A Switch)

Alonzo · 5453

0 Members and 1 Guest are viewing this topic.

Offline Alonzo

  • Sr. Member
  • ****
    • Posts: 461
  • Not all those who wander are lost
on: February 09, 2012, 08:02:51 PM
So I have a number of DIY amps that are growing on me and may stay around awhile.  Since my power tube complement is limited, I was thinking of adding either a soft start circuit or switch to keep my tubes healthy and happy.  My first inclination is to put a switch in and a sticky of "1st, count to 10, 2nd" on each amp but I wanted experienced opinions before proceeding.  I'd like something that fits easily across 2A3, 45, el84 and a KT88 circuits.  So to me this would either be an easily configured board or a simple switch to delay startup.  Any opinions on what's better?

Alonzo
Gameroom:>Mainline to HD820, SR45 to Pipette
>BeePree Kaiju & SII to Altec 19 knockoffs
Office:>BH Stat amp to Koss 95x, T20 SET to JBL 4309s
Den:> MorePlay 845 SET to Altec Valencia's


Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5833
Reply #1 on: February 09, 2012, 10:43:53 PM
All of the reasonably simple solutions are compromises, and all of the ideal solutions are circuit dependent. They also depend on what you want from them - longer tube life, better sound sooner, reduced turn-on and turn-off thumps, etc.

For example, the "soft-start" board in Paramount delays and slowly applies plate current to the driver tube. This helps protect the 2A3 if it is configured for the direct coupled 2A3 circuit, by keeping the grid near ground. It does not prevent high voltage on the output tube before the cathode warms up, so the longevity could be even greater. Amps which are not direct coupled will see little advantage, and it won't survive unless the voltage is limited at startup - which the Zener bypass on the shunt regulator provides, until the shunt reg tube warms up and takes over. Other amps that do not use C4S plate loads on the driver can't use it, and without regulation they may even destroy it.

For another example, again Paramount, ideally the output transformer would be shorted across the primary until the tubes are warm and the voltages stable; this would prevent the start-up transient from magnetizing the core. If you have nickel core transformers, this eliminates the 1-hour warmup with music needed to de-magnetize the transformer for optimal performance. This is only applicable to parafeed outputs, and mostly just to nickel core transformers.

The most common approach for most circuits is to delay and slowly ramp up the high voltage. Switches and time delay relays can delay the voltage and eliminate initial over-voltage, but not slow the rise. Slow-warmup tube rectifiers can do both, but the rectifier warmup must be slower than the slowest tube, and the rectifier itself is not protected.

In the case of switches and relays, there are no affordable switches that are rated for more than 240vAC. (DC ratings are much less and only AC should be switched.) Switches are widely used beyond their ratings, but of course in that case are less than reliable and potentially dangerous. (Paramount, again as the example I am most familiar with, has a 200v winding with voltage doubler, so it can be switched.)

The cleverest simple scheme I know requires separate transformers for heater and high voltage power. A pair of double-pole switches are wired in parallel for the heater power, and series for the high voltage. This way you can throw either switch and get the heaters going, then throw the other switch to start the high voltage power.

Paul Joppa


Offline xcortes

  • Sr. Member
  • ****
    • Posts: 997
Reply #2 on: February 10, 2012, 03:37:02 AM
Quote
For another example, again Paramount, ideally the output transformer would be shorted across the primary until the tubes are warm and the voltages stable; this would prevent the start-up transient from magnetizing the core. If you have nickel core transformers, this eliminates the 1-hour warmup with music needed to de-magnetize the transformer for optimal performance. This is only applicable to parafeed outputs, and mostly just to nickel core transformers.


Paul, this has been on my shortlist for a while (Nickel opts here). Most times I don't care about the one hr (I'd say at half hr you're 90% there) wait because I usually play some background music while dining before any serious listening. But sometimes I just skip listening to music because I can't/don't want to wait.

Anyway, any special requirements for the switch? I'd assume that, since there's no DC at the primaries any decent switch would be ok?

Thanks

Xavier Cortes


Offline Grainger49

  • Hero Member
  • *****
    • Posts: 7175
Reply #3 on: February 10, 2012, 05:07:46 AM
I would be interested in PJ's response as well.  I have a couple of spring return to off switches I ordered by mistake years ago.  Sounds like the perfect thing.



Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5833
Reply #4 on: February 10, 2012, 06:28:25 AM
Yes, I think there are no severe requirements on that switch.

I did make a board once with (IIRC) three delay relays with different timings, but it was not a success - the high voltage initial surge reset all the other relays. I never got back to it for a more careful design, partly because there's no good place for yet another PC board, and partly because while I'm at it I would like to make the DC relay supply also provide regulated DC for the driver heater - now the board is getting pretty big!

Maybe someday I'll make an "ultimate Paramount" on a larger chassis with all the delay relays and all film caps. Probably cost $10K though - so maybe not!

Paul Joppa


Offline Grainger49

  • Hero Member
  • *****
    • Posts: 7175
Reply #5 on: February 10, 2012, 07:56:44 AM
Paul,

Sounds like a two chassis per channel amp.

And the spring return DPDT toggle switches will go into the Paramours soon.  I will only use them for immediate serious listening, like Xavier.
« Last Edit: February 12, 2012, 06:20:23 AM by Grainger49 »



Offline xcortes

  • Sr. Member
  • ****
    • Posts: 997
Reply #6 on: February 10, 2012, 11:11:25 AM
How about using this circuit (with a wallwart) to short the opt for a few seconds?

http://www.circuitdb.com/circuits/id/92

http://www.circuitdb.com/circuits/id/25

Xavier Cortes


Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5833
Reply #7 on: February 10, 2012, 03:05:19 PM
Sure, there are many time delay circuits around. You'll have to connect the wart to the power transformer primary, and find a place for it. Make sure it's a transformer wall wart, not a switching type. And find a place for it and for the board with the relay circuit.

I think it would be easier to make a FW bridge/cap power supply from the 6.3v winding, putting that PSU on the same PC board as the relay and delay circuit. Just use a 3.3-v zener in place of the B-E junction of a transistor. I'd go for a longer delay time, but that's just me.

If you do that and have the soft-start, put in another relay (or use a DPDT relay) to un-short the delay capacitor on the soft-start board. The function there is to discharge that cap immediately so you don't have to wait 20 minutes. Replace the 33K on the SS board with the appropriate relay connections.

Paul Joppa


Offline Alonzo

  • Sr. Member
  • ****
    • Posts: 461
  • Not all those who wander are lost
Reply #8 on: February 10, 2012, 03:25:39 PM
This is getting way above my paygrade in circuit knowledge, I'll go study up but for the specific problem of extending tube life and getting the best sound out of each upgrade would this be a good compromise:  seperate htr/pwr transformers with the double pole switches mentioned before, with a seperately switched reticifer to prevent the instant high power to the output tube and a start sequence of htrs (all), reticifer power, output power.  I would use the rectifier as a sacrificial anode, but since everything affects the output sound would this be the wrong way to go (slowly dying reticifer degrading the output tube power and it's sound).?

Alonzo
Gameroom:>Mainline to HD820, SR45 to Pipette
>BeePree Kaiju & SII to Altec 19 knockoffs
Office:>BH Stat amp to Koss 95x, T20 SET to JBL 4309s
Den:> MorePlay 845 SET to Altec Valencia's


Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5833
Reply #9 on: February 10, 2012, 08:36:58 PM
Of course, I like Schottky diodes, so take my comments with your own spin. But my feeling is that rectifier tubes are relatively inexpensive and are built for abuse, and they have the advantage of a slow rise in voltage - especially if they are indirectly heated. If you have tube rectifiers, I'd fire them up along with the HV.

There is yet another method - I should have posted but my mind was not on tube rectifiers and I forgot. You can switch the rectifier heater power after the main power, without needing separate transformers and without needing high-voltage switches. This gives a delayed start AND a slow ramp in voltage. The rectifier will suffer, but it's built for that.

Another point that just occurred to me - if an audio tube loses 30% of its emission, it's dead. But the rectifier tube only eats up some 10% of the voltage; if that rises to 13% it's not that big a sonic loss.

Interesting discussion.

Paul Joppa


Offline Alonzo

  • Sr. Member
  • ****
    • Posts: 461
  • Not all those who wander are lost
Reply #10 on: February 10, 2012, 09:12:37 PM
Now this is more up my alley, experimentation without full comprehension but plenty of fire extinguishers...So for what you mentioned, using a single transformer, both heaters would be switched together output and rectifier with power to outputs also, with the rectifier power being on a seperate switch, turned after the heaters are on.  This would light the heaters, and once the switch is thrown for the rectifier, provide power ramp for outputs.  How would this be affected by chokes or filters?  Sorry for the beginner questions, I'm talking this out in my head as I type, it helps me understand.

Alonzo
Gameroom:>Mainline to HD820, SR45 to Pipette
>BeePree Kaiju & SII to Altec 19 knockoffs
Office:>BH Stat amp to Koss 95x, T20 SET to JBL 4309s
Den:> MorePlay 845 SET to Altec Valencia's


Offline RayP

  • Sr. Member
  • ****
    • Posts: 301
Reply #11 on: February 11, 2012, 02:10:25 AM
The link shows the dual switch arrangement. I've been using this arrangement with Radio Shack DPDT switches and a RS low voltage transformer for years. I also find it useful when first testing a new amp. I can check that the heaters are all working correctly before I turn on the high voltage.

There is also a school of thought that having a separate low voltage transformer for the heaters is advantageous to the sound quality of the amp.

http://829b.com/dualdpdt.aspx

I'm curious about the nickel transformer demagnetizing period. What changes in the sound over the hour period that you can notice. Bass?

ray

Ray Perry


Offline Jim R.

  • Hero Member
  • *****
    • Posts: 2194
  • Blind Bottlehead
Reply #12 on: February 11, 2012, 05:50:03 AM
Another possibility is to use SS rectifiers on the outer windings and put a damper diodon the center tap, but of course you have to have a center tapped PT i order to be able to do this.  Then you can use a directly heated diode or a switch for the filament power.

Of course I just woke up so not sure if I'm making sense :-).

-- Jim

Jim Rebman -- recovering audiophile

Equitech balanced power; uRendu, USB processor -> Musette DAC -> 5670 tube buffer -> Finale Audio F138 FFX -> Cain and Cain Abbys near-field).

s.e.x. 2.1 under construction.  Want list: Stereomour II

All ICs homemade (speaker and power next)


Offline Paul Joppa

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 5833
Reply #13 on: February 11, 2012, 06:42:01 AM
Now this is more up my alley, experimentation without full comprehension but plenty of fire extinguishers...So for what you mentioned, using a single transformer, both heaters would be switched together output and rectifier with power to outputs also, with the rectifier power being on a seperate switch, turned after the heaters are on.  This would light the heaters, and once the switch is thrown for the rectifier, provide power ramp for outputs.  How would this be affected by chokes or filters?  Sorry for the beginner questions, I'm talking this out in my head as I type, it helps me understand.
Sorry, I was not very clear. (I also can't quite figure out exactly what you said ... it's very difficult to be clear with just words!) I will try again:

The idea is to put the second switch in the rectifier filament (heater) circuit. When only the transformer (first switch) is on, all heaters except the rectifier are on, and the rectifier plates have high voltage on them but no high voltage appears at the rectifier output becaus it is not conducting yet. After 30 seconds, switch on the rectifier heater, and as it heats up the high voltage will ramp up.

Because there are serious safety issues with high voltage power supplies, I urge you to delay experimentation until you feel you have full comprehension. Then when experimenting, start with the assumption that you didn't actually comprehend, and expect the unexpected.


Paul Joppa


Offline Alonzo

  • Sr. Member
  • ****
    • Posts: 461
  • Not all those who wander are lost
Reply #14 on: February 11, 2012, 07:18:50 AM
Sorry I wasnt clear.  I understand your explanation well enough to give it a try.  Also the link RayP posted has a good clear drawing just using a double switch arrangement. 

What traits make Schotty diodes your preference over rectifiers?
« Last Edit: February 11, 2012, 07:21:35 AM by Alonzo »

Alonzo
Gameroom:>Mainline to HD820, SR45 to Pipette
>BeePree Kaiju & SII to Altec 19 knockoffs
Office:>BH Stat amp to Koss 95x, T20 SET to JBL 4309s
Den:> MorePlay 845 SET to Altec Valencia's