EL156 to KT120 bias

[Willabe]

I'm not sure how to go about this?

I have 2 Eastern Electric M156 mono blocks that the owner want's me to change the EL156 tubes to KT120 tubes.

They have a separate bias adjust pot on the face panel for each tube socket. And there's also very large dual needle bias current/power output meter on the front panel. 

He did not give me the EL156 tubes to see where they were biased at.

Schematic (attached below) lists the PT power tube B+ wind as 510vac/400mA. The choke lists as 450mH/350mA.

I talked with 1 of the partners how had these amps built, but he was not the design engineer. He said they were having problems with the new EL156 power tubes and decided to go with new KT120's as replacements. To do this he thinks they changed 1st bias dropping series R (R31) from2K/2w to 100ohm/2w. 

Now this is a mono block for stereo, so given that and the PT/choke ratings, what should I set the current draw for on the new KT120's?

Here's new EL156 tube specs;

    Filament Voltage = 6.3 V
    Filament Current = 1.9 A
    Plate Voltage = 440 V
    Plate Current = 100 mA
    Screen Grid Voltage = 350 V
    Screen Grid Current = 16 mA
    Control Grid Voltage = S-bias
    Load Resistance = 4000 ohms
    Transconductance = 13.0 mA/V
    Maximum Data:
    Plate Voltage = 800 V
    Screen Grid Voltage = 450 V
    Plate Power = 50 W
    Screen Grid Power = 8 W
    Cathode Current = 180 mA

Here's the link for the KT120 specs.   

http://tungsol.com/tungsol/specs/kt120-tung-sol-specs-curves.pdf

[shooter]

check the -G1 volts without tubes, figure out the range next.

from the data sheets and guessing, ~~480plate, something ~~ -20 to -25 will be "safe enough" as a starting point so you can get an actual current value, then crank the math for say ~~~60% max plate

[Willabe]

Gut shot.

The mono block weighs 62.2 lbs.

[Willabe]

Ok, I took some readings.

R31 has already been changed to 100ohm/2w from 2K/2w.

At the junction of R31, C9 and D3, raw -bias dcv =   -41.7dcv

At pin 5 on all 4 KT120's at max -dcv =   -40.5dcv, min.   -27.8dcv

At the max -bias, the front panel meter shows the tubes soar to ~90mA in a few seconds and are still climbing when I turn the amp off.   

I then looked for red spots on the KT120 plates, I see none. But I have to do this very quickly. Doing this the front panel current meter was pined, 110mA.

Rob Robinett's bias calculator shows;

KT120/60w, AB fixed bias;  50% = 47.2mA, 60% = 56.7 mA, 70% = 66.1mA

To double check the panel meter, I measured across the 10ohm K R on the power tubes, 1.47dcv.   

And the B+ dropped to 635dcv, unloaded 726dcv and climbing at turn off. 

The -bias wind is a dedicated 2 wire wind, not a tap off the B+ wind.

It's wired as a 1/2 wave, I think it needs a voltage doubler or needs to be a FWB to double the raw -bias dcv?

[shooter]

measured across the 10ohm

were you at -40vdc when you did this?

Rob Robinett's bias calculator shows;

you got the #'s based on 635vdc plate?

doing E * I comes up with 635 * .147A = 93W but I'm not a fixed bias PP guy so that's probably a bad # 

[Willabe]

measured across the 10ohm

were you at -40vdc when you did this?

Yes.

Had to, it's the coldest setting.

Rob Robinett's bias calculator shows;

you got the #'s based on 635vdc plate?

Yes.

[shooter]

on the 120 data sheet (1st graph)if you extend the lines out to 650vdc
looks like you need a bias ~~ -65 - 70 but you'll still be in the 120 140mA Ia range

If you look at the max section for 800vdc, you need -200vdc for G1 

[Willabe]

on the 120 data sheet (1st graph)if you extend the lines out to 650vdc
looks like you need a bias ~~ -65 - 70 but you'll still be in the 120 140mA Ia range
If you look at the max section for 800vdc, you need -200vdc for G1 

Schematic (attached below) lists the PT power tube B+ wind as 510vac/400mA. The choke lists as 450mH/350mA.

Well, PT B+1 wind is good for 400mA, but the choke is only 350mA.

Screens and the 2 @ 12AU7's have separate B+ supplies.(There's 2 B+ supplies in this amp, B+1, and B+2/B+3.)

The -bias wind is a dedicated 2 wire wind, not a tap off the B+ wind.

It's wired as a 1/2 wave, I think it needs a voltage doubler or needs to be a FWB to double the raw -bias dcv?

[2deaf]

The -bias wind is a dedicated 2 wire wind, not a tap off the B+ wind.
It's wired as a 1/2 wave, will a full wave double the raw -dcv voltage? Or do I need to go to a FWB to double the raw -bias dcv?

FWB will only get you a couple of volts.  You will probably need to go voltage doubler like the attached.  R31 can be adjusted for different bias ranges up to -80V or so.  The doubler will have more ripple than the original.  If this is a problem, you can add C_o for more filtering.

[Willabe]

Thanks 2deaf, I think these KT120's in this amp will need every bit of -80dcv.



There's a 100ohm R in there now. I think I'll just leave that in for now to get the max -dcv 1st and see if that tames the KT120's?

[Willabe]

PT B+1 wind is good for 400mA, but the choke is only 350mA.

Screens and the 2 @ 12AU7's have separate B+ supplies.(There's 2 B+ supplies in this amp, B+1, and B+2/B+3.)

Question is with this iron, how low do I have to set the bias current?

The front panel power meter only goes to 160w. I don't think they intended for it to be run full out at 160w?

[2deaf]

There's a 100ohm R in there now. I think I'll just leave that in for now to get the max -dcv 1st and see if that tames the KT120's?

The 4.7K for R31 should give you a range of -35 to -50 (on paper with some approximations) and 100R for R31 should give -75 to -47.  Starting with the coldest first sounds like a good idea to me.

[shooter]

Look up the hourly rate for a freelance engineer 

Here's the range I get, visual guessing 

[shooter]

current meter was pined, 110mA.

You might have to re-scale 

[shooter]

Couple other ideas, usefulness aside 

Put a 100 ohm 25W R ‘tween C15/16  my #’s say 40vdc drop using 400mA
Increase the self biased to 20ohms from 10

[Willabe]

Look up the hourly rate for a freelance engineer 

Thanks shooter. 

I'm going to try and get ~ -80dcv and see how much it tames the KT120's.

[DummyLoad]

triode curves from -15Vg1 to -30Vg1 with anode at about 200V with Vg1 -15V; 310V with Vg1 -30V both with Ia at 90mA. mu is about 7.33


based on rule of thumb Vg2/mu = -(Vbias) ;  so about -60V?


--pete

[DummyLoad]

bias PS sim.

range -53V to -70V


--pete

[Willabe]

based on rule of thumb Vg2/mu = -(Vbias) ;  so about -60V?

Thanks DL.   

How did you come up with the mu#?

[Willabe]

I disconnected the PT -bias leads and gator clipped in a 120vac PT, controlled it's output with a variac. 

Started with the raw -bias at -80vdc, with all 4 -bias adjustments pots set for max -bias dcv. The tubes were in shut off. Went to -70vdc, coldest tube was at ~15mA. Hottest tube just under 35mA. That's according to the front panel meter.

Then I set the raw -bias to ~50vdc, and I was able to set all 4 tubes at 50mA each with the 4 -bias adjust pots. B+ plate dcv was ~695.

So, I think I'll make a small turret board to build that voltage doubler circuit on and abandon the amps PCB circuit.

Should work out fine. 

I'll post as I go. I'll have to order some lytics for the new doubler circuit.

Thanks for the help guys.   

[Willabe]

bias PS sim.

Is this sim a free online program?

[shooter]

 

the math shows ~ 58% max plate using 50mA and 695vdc
unless you got a crazy hot signal you should stay in undistorted range quite well

[DummyLoad]

bias PS sim.

Is this sim a free online program?

kind of still free, but in limbo - mouser had/has multisim blue as a give-away in hopes that you'd buy the $300.00 annual subscription. i guess it didn't pan out. the subscription is no longer offered, and neither are any of the perpetual licensed packages they used to list in their catalog... i guess national instruments and mouser are parting ways? multisim blue free version doesn't allow edit/create new parts. you can edit an existing part and change the SPICE parameters is how i work around that - that and has a limited parts library.   

multi-sim blue still works and gets updates, but i imagine it will stop receiving updates when the next major release is published.

https://www.mouser.com/multisimblue/support/downloads/download-multisim    - that link takes you to the download page.


--pete

[DummyLoad]

based on rule of thumb Vg2/mu = -(Vbias) ;  so about -60V?

Thanks DL.   

How did you come up with the mu#?

from the triode mode curves. mu = deltaVanode/deltaVg1 - i drew a horizontal line across 90mA. noted the value of Vanode at -15V Vg1 and Vanode at-30V Vg1 - that's the delta of Vanode with a delta of Vg1 from -15V to -30V - my eyes see Vanode of ~200V @ 90mA at -15V Vg1 and Vanode of ~310V at 90mA at -30V Vg1


so - 310-200 110V / (-15) - (-30) =  110V / 15V = mu = 7.33 

why did i pick 90mA for my line? b/c -15V g1 curve intersected nicely at 200V Va + it's in what looks like it's in an area that's the most linear part of the curves. realistically i should have picked Vg1 curves around the 600V mark and current line of what we'd be biasing at, but the Vg1 curves stop with Va at 500V. interesting since the MAX Vanode in triode mode is stated to be 650V (max Va in triode mode is usually around max Vg2 rating). 



--pete

[silverfox]

Paging       SLUCKEY;      PRR

[shooter]

 
We got this, let 'em rest 

[DummyLoad]

brad, since these are HI-FI - if you have an o-scope, then check for x-over distortion with small signal output and full power output if possible.  e.g., tune bias for minimum THD.


--pete

[Willabe]

The owner has 3 sets of these and I went over to his home last night and opened another of these mono blocks up.

I measured the -bias wind and got ~42acv. IIRC, the 1 I have open here I measured ~30acv. So I don't know if they under wound PT -bias wind or if they did a 2nd run of these PT's and upped the -bias wind acv a little?

We adjusted the -bias on his last night and the coldest we could get the KT120's was 55mA. His has a max -bias of ~52dcv on pins 5 of the tube sockets.

I have to open up the other mono block I have here and measure it's -bias winds acv.

He found the manual, and it says to bias the tubes, EL156, at 40mA each. Manual claims 160w RMS output on a single mono block with EL156, at 40mA each.

While I was there I had him turn up the pair of mono blocks through a big/huge set of speakers and we got it up to 120w peaks on the front panel meter, sounded pretty good. And he had separate self powered subs hooked up with it.   

[shooter]

Manual claims 160w RMS

I wouldn't worry too much about subtle nuances in the music but if the client does, call me I got all kinda cool things for sale 

[Willabe]

Manual claims 160w RMS

I wouldn't worry too much about subtle nuances in the music but if the client does, call me I got all kinda cool things for sale 

 

[Willabe]

They gave me another EL156 mono block that's blowing fuses.

I tracked the problem down to 2 blown B+ rectifier diodes, D4/D5, of the FWB power tubes plates supply.

The B+ filter caps just after the FWB seem to be fine.

What would have killed these 2 diodes? There 1000v/3A. 

[shooter]

1000v/3A. 

those 2 specs are what kills diodes, exceeding PIV is more likely in a low current system, seconds is usually all it takes.
excessive current takes longer.  diodes typically fail "short" till they open, which means AC gets into the DC side for a short time, again usually seconds.

[Willabe]

In a FWB 2 diodes on each side are in series, so that would be 2000v PIV/3A. 

That should be more than enough. 

[shooter]

2 diodes on each side are in series

my brain might still be froze but I don't see 2 in series;
during the pos 1/2 cycle D4 & D7 conduct, and D6 & 5 block, during the neg cycle they flip, 6&5 conduct and 4&7 block.
it's still teens with windchill and I'm still shivering so I might be out to lunch 

[sluckey]

during the pos 1/2 cycle D4 & D7 conduct

If you draw that out... top side of PT secondary to D4, to one side of load, thru the load, on to D7, then back to the bottom side of PT secondary... you'll see a nice series circuit. True, the diodes are not touching but they are in series.

[shooter]

they are in series.

ok so I got the pairs correct 
but if they did blow because PIV, there probably wasn't any, or much current, just voltage potential.  So if D4 "sees" >+1000vdc, won't D7 still be "sleeping" til there's enough current to wake it up 

[PRR]

For 520V AC, this configuration, the peak inverse voltage on each diode is 720V.

(The inverse is per diode, from reverse swing to charged-up capacitor.)

_I_ like a 2X margin on PIV, and more if I can buy it cheap. So in my book this is inadequate. However it is not wrong.

The cold-start current may approach 50 Amp. The running current is a hair over 1 Amp. I like more margin here, especially in a hot chassis, but it is not wrong.

Your bet. Call it a fluke, replace same-type, and see if it happens again. Or go conservative, find much higher rated diodes (and this will be non-trivial added cost, though maybe less than the cost of another bench-time).

[Willabe]

For 520V AC, this configuration, the peak inverse voltage on each diode is 720V.

_I_ like a 2X margin on PIV, and more if I can buy it cheap.

So at least 1500v PIV rated diode?

The cold-start current may approach 50 Amp. The running current is a hair over 1 Amp. I like more margin here, especially in a hot chassis, but it is not wrong.

So each of the 4 diodes see 'a hair over 1A running current'? Not 1A/4.

3A rating is almost 3X running current, not enough? 

The mono block chassis, tube sit up right in free air space, no cover and all tube sockets are vented, bottom chassis plate cover is well vented too.   

Or go conservative, find much higher rated diodes ...........

Like, 1500PIV/4A or 1500PIV/5A?

[shooter]

Your bet. Call it a fluke, replace same-type,

I think of diodes as expendable also, rather blow a diode than 4 tubes because bias was 0 and max current is running raiment. but it is a fine line who wins the smoke test!

[DummyLoad]

these are what i would consider for the application. 1.2KV + 120A surge.

https://www.mouser.com/ProductDetail/IXYS/DS2-12A?qs=sGAEpiMZZMtoHjESLttvkpfqOErzsE885m0CkeC9TNc%3d

https://www.mouser.com/datasheet/2/205/DS2_DSA2-313088.pdf

1.2KV units should suffice - non trivial price tag tho - $6.60 qty. 1 - oouch! lower cost alternative are available if you could/would be willing to use TO220 or TO247 package. with TO220/TO247 packages, you'd need to drill holes in chassis along with thermal grease, insulation mica, screws nuts, etc..

https://www.mouser.com/ProductDetail/Infineon-Technologies/IDP30E120?qs=sGAEpiMZZMtoHjESLttvkmfZcEXpcVA5RrKTS9b0c9o%3d

https://www.mouser.com/datasheet/2/196/IDP30E120_v2_3G[1]-78909.pdf   <<< copy to browser location - the [1] in-line in the link is breaking things. forum software does not like brackets in links.

--pete

[PRR]

Another plan is cheaper 1000V diodes, two series per leg. Mouser RURP8100 is ten for 9 bucks. Nominal 1KV 8A. TO220 package. At 1A RMS it "should" run cool enough to not need a heatsink. If you do need sinking then it must all be insulated.

> copy to browser location - the [1] in-line in the link is breaking things.

Yeah, that [] in the url pooches the forum's parsing. Here is a tinyURL link for folks without proper mice:
https://tinyurl.com/Hoffman-IDP30E120
Or- mouser.com and Search for IDP30E120

[Willabe]

Thank you PRR, Pete. 

[Willabe]

I've installed the -bias voltage doubler in 4 mono blocks for use with KT120 tubes.

They work fine. Had to find the right value for the series dropping R.

On 2 of the mono blocks, the bias wind was ~32vac. On the other 2 mono blocks, the bias wind was ~42acv.

With the 32acv wind, 1K got 40mA at the center of the bias pots sweep. With the 42acv, needed 4K7 to center the 40mA current draw on the bias pots swing.