Tube amp with KT120's, 150 watts!!

[pylet2000]

I posted something on this before, but wanted to start a clean thread that just focuses on the amp.

I designed it around a Fender Twin Reverb, roughly. It's a simple circuit, volume / treble / mid / bass. LTP phase inverter like the Fender. The plates run at 650vdc, the screens are at 600.

It sounds a lot like a Blackface Fender, but deeper. There's a "roundness" to the sound that follows the notes all the way up the neck. In comparison to a Mesa 400+, it has more character to it, feeling more like an instrument amp and less like a broadcasting station, if that makes any sense. It's more interactive and rewarding to play than the 400+, weighs about 1/3 as much, but of course won't make the kind of power the 400+ makes.

I figure it's got roughly $600 wrapped up in it, not counting the preamp tubes which I had lying about. The Mercury Magnetics OT was a pricey sucker!! Anything worth doing is worth doing right though, and a tube bass amp may have no more single important component that the OT, not counting the speaker / cab. The PT is a Hammond 370KX, using a bridge rectifier to get the ridiculous B+ I wanted. It runs HOT. I suspect the first filter cap stage is part of that. I use 10 total 220 uf / 350v caps in series / parallel configuration, giving an entire 220uf / 700v section to the first stage before a 20 ohm 5 watt resistor, then an additional 220uf filtering for the plates. The screens get 110uf, then things go back to sane Fender levels for the PI and preamp. It all modeled out really nicely in the Duncan power design software / freeware that's out there. Not a hint of ripple. The bridge rectifier required 8 diodes (I used the UF4007). I tried the traditional 4 diode configuration, but after two failed diodes it became apparent that a series pair was required for each leg.

Bias with these tubes is actually something to be concerned about. There's a real good chance that many amps that use KT88's might not be able to bias the KT120's into compliance. It requires about -65vdc, and that's at the full 60 watts dissipation per tube. Your regular old bias tap isn't going to cut it. I used the 5V filament winding into the speaker side of a Champ style OT, giving me about -90vdc when rectified half wave. I also used a separate bias supply pot for each tube, all the KT88 data sheets call for this and it seemed prudent here too.

I'm still waiting on my chrome barrel knobs to show up, and forgive my phone camera, what awful pics it takes...

[pylet2000]

And here's a pic of the back.

I think I'm going to try and record some audio clips today, too.

[majortoby]

Beautiful.

[pylet2000]

Originally Posted by majortoby

Beautiful.

Thanks!

[JMac4strngr]

Really nice job. What style/direction are You gonna go for in a head enclosure?

[B-string]

Nice job! But yes too much filter before the resistor, transformers do not operate well with that much capacitance on the secondary. Do you have any bleeder resistors in the HV supply?

[pylet2000]

I'm really aiming to make this a fliptop. Gotta show that glass!

I'm leaning towards a fEARful 15/6, there appears to be enough room behind the mid enclosure to make a flip panel. The amp itself measures 17" W x 8" H x 6" D, so it will be tight, but doable.

[pylet2000]

Originally Posted by B-string

Nice job! But yes too much filter before the resistor, transformers do not operate well with that much capacitance on the secondary. Do you have any bleeder resistors in the HV supply?

Thanks!

Dropping the first filter stage to 110uf is on my list of things to try, perhaps I should get to that sooner than anything else.

I didn't put any bleeder resistors in, I know it's generally good practice to help ensure the caps will discharge when unpowered. That's also on the "to try" list. As it is, the caps will discharge down to roughly 6v on their own, but it takes about 30 seconds.

[BassmanPaul]

Originally Posted by pylet2000

I didn't put any bleeder resistors in, I know it's generally good practice to help ensure the caps will discharge when unpowered. That's also on the "to try" list. As it is, the caps will discharge down to roughly 6v on their own, but it takes about 30 seconds.

No - the point of the bleeder resistors is to balance the voltage across the series connected caps. The fact that they also eventually discharge the caps is incidental.

[B-string]

Originally Posted by pylet2000

Thanks!

Dropping the first filter stage to 110uf is on my list of things to try, perhaps I should get to that sooner than anything else.

I didn't put any bleeder resistors in, I know it's generally good practice to help ensure the caps will discharge when unpowered. That's also on the "to try" list. As it is, the caps will discharge down to roughly 6v on their own, but it takes about 30 seconds.

Bleeders also improve filtering, dropping down to 110uf is a good idea! The caps should not bleed down on their own? Where is load they are bleeding through? Might be worth exploring now.

[pylet2000]

Originally Posted by B-string

Bleeders also improve filtering, dropping down to 110uf is a good idea! The caps should not bleed down on their own? Where is load they are bleeding through? Might be worth exploring now.

Thanks again, you guys rock.

Just dropped the first filter stage to 110uf. No audible difference, so that's just a good all around move. Thanks for confirming my hunch. I remember reading why larger first stage filter caps are harder on the PT, but it escapes me now. Something like "the bigger the cap, the bigger the spike to PT sees." I also added 220k bleeder resistors across the caps on the first filter stage. No difference in running voltage or change in discharge rate when powered off.

My understanding on cap discharge, and I'm always willing to further my knowledge so chime in if I'm off base, is that it's the tubes that are causing the discharge. They are still hot and still drawing idle current even when the amp is "off", and will do so as long as they have even a little voltage available to them and still have their heat. Thus the old adage of killing the power but leaving the amp in standby and playing a note to drain the caps. I don't think that's necessary, I think they discharge anyway, and as long as all the caps are after the standby switch that wouldn't matter either. If you leave the tubes out and turn the amp on, that's a whole different story, those caps will hold their charge until they find a way to get rid of it, very dangerous condition, unless you have the bleeder resistors in place.

But you did cause me to question my logic, so I double checked that, without the bleeder resistors, there is no path to ground on the high voltage supply. Always good to confirm.

I hadn't considered there being a need to balance the voltage across the series filter caps, but it makes sense I suppose. Thanks for pointing that out. I'm going to guess that every series pair of caps needs it's own pair of bleeder resistors for the purpose of balancing voltage across them?

[lneal]

Nice work. There's nothing like the feeling of building your own gear, especially when it works like you want it to and is useful. Several years ago I drew up plans for a 300 watt all tube, transformer coupled power amp that uses (6) 6146W output tubes, but life being what it is the time just isn't there.

Check out my sig...............

[BassmanPaul]

Originally Posted by pylet2000

I'm going to guess that every series pair of caps needs it's own pair of bleeder resistors for the purpose of balancing voltage across them?

You guess correctly.

BTW the 220KΩ resistors should be sized at 2 watts.

Now we need to see an under chassis photo.

[1958Bassman]

Any chance of you putting the audio clips from the preamp build online again?

[B-string]

Originally Posted by BassmanPaul

You guess correctly.

BTW the 220KΩ resistors should be sized at 2 watts.

Now we need to see an under chassis photo.

You dirty old man ......yes, lets see the amp porn!

[MAMMOTHvolume]

just curious why you decided to run the kt120's at the front? did you have any stability problems or have to do any funky lead dress?

I think the reason you're PT is getting to hot is that it doesn't supply enough current for the HT for the kt120's. They need about 440ma and your PT is doing 360ma. You should have no problem with huge caps as long as your not rocking a valve rectifier.

[BassmanPaul]

Originally Posted by B-string

You dirty old man ......yes, lets see the amp porn!

Hey! Less of the old!!

[B-string]

Originally Posted by MAMMOTHvolume

just curious why you decided to run the kt120's at the front? did you have any stability problems or have to do any funky lead dress?

I think the reason you're PT is getting to hot is that it doesn't supply enough current for the HT for the kt120's. They need about 440ma and your PT is doing 360ma. You should have no problem with huge caps as long as your not rocking a valve rectifier.

I disagree, yes tube rectifiers will arc and destruct with excessive cap loads. But I have replaced to many PTs from guys "improving" their amps by over capping the plate supply right after the rectifier bridge.
The available plate current you pointed out could be a problem to obtain full output though.

[B-string]

Originally Posted by BassmanPaul

Hey! Less of the old!!

Sorry Paul But, you do know a couple more rocks before they became dirt than myself?
We are not old....just well seasoned

[Eight_Stringer]

Nice work pylet2000, even more so given your learning as you go. Had to look up what a KT120 was, all news to me.

The lads are correct about the voltage equalisation resistors around any series connected filter capies. Prolong service life during start up phase ( Electrolytic capies are +100% -20% tolerances unless stated otherwise ). Dependent on the way you rectify the Plate HT, a bridge configuration without series choke will load the AC secondary HT current at about 150% of the plate current maximum, the bigger the capies the greater the load on the PT secondary HT winding, eventually it is counter productive with this high current demand causing output voltage regulation problems. Less is more at times. Think you have a handle on that concept already, well done you.

Do be careful of potentiometer derived control grid bias setting. If you must use pots, put the pot on the "earthy" or cold end of the negative bias drive circuit, this was mentioned by another knowledgeable TB person also. That way if the pot goes open the control grid will go high negative potential and protect the tube. Also you need to address "hard or stiff" bias supply design ie low source impedance for DC.

High plate HT has the advantage in that power is available at a rate of P = E squared over load R. That is for each volt the ht increases the power is squared. Simple example, 1 volt squared times 1 amp is 1 watt, 2volts squared times 1 amp 2 watts, however that is for a fixed current. Since if voltage doubles and a fixed resistance load is used, say 1 Ohm then the current will double also and you get a square law for the power produced or consumed, 2 volts squared at 1 ohm load will yield 2 amps, P=EI and you get 2 time 2 = 4 Watts Why bridged amps yield a big gain in power output, and the careful attention to the load resistance in that mode is required.

Use 700VHT on the quad EL34's here for my main amps. Rarely do you get something for nothing in electronics, and flashover has to be a concern when using elevated HT's.

All the best with your project, many hours well invested, playing and technical skills alike.