kt 120

[Andy Le Blanc]

Stumbled over a KT120, the new tungsol label, kinda hoping it holds up to its published spec.
60w plate dissipation, a pair of these and twin PT might be like peas and carrots.

[es345]

I have done an evaluation powerstage with that tubes for a bass amp . You can get a maximum of 200W clean power out a pair but then you are really at the limits. For a guitar amp using heavy overdrive I would not go over 120-130 W. I have attached the evaluation circuit. Don't be astonished about the PI, I just wanted to test the EF86 for such an application.

[Colossal]

Nice Andy, I was kinda wondering when you'd get around to the KT120 in your lab. Looking forward to seeing what you come up with.

BTW, I'm calling your screen resistor mod the LeBlanc Screensaver

[Andy Le Blanc]

I really like kt88's, I've built around kt90 and kt100 in the past.
It becomes a different design consideration when the tube spec exceed the capability of your power supply.
You have look at the power supply differently, your not going to in any way be able to meet the max plate dissipation figure,
so have to have decide where to draw the line based on the transformer spec.
The last build with kt 90 was a twin pt and 100w jcm opt, I mis-matched the speaker to get the plate load I was after, work really well.
Other than that it was a very simple rig, and that brings me to another observation,
something happens when you put a simple garage sale preamp and inverter onto a power-side with significant muscle.

The bench sketch of the premier 88n circuit fell out from under a pile of mess from under my desk the other day, that looks like it might be one hell of a rig with real umff behind it.
Zero bias pre, split load inverter and driver stage, damn close to a danelectro.

I appreciate the kudos there, I glad its useful for fellow DIYer's.

[JazzGuitarGimp]

Hi Andy,

At least for a prototype, you could buy two identical PT's and either parallel them for more current, or series them for more voltage. Once you have a working prototype, you could have a custom transformer wound for the amp. Another thought: maybe some of the transformer companies have a replacement for the 300W SVT. Perhaps that would be a less expensive way to get there than having a custom part made.

I forget how I discovered the KT120, but it was about 18 months ago. I may have been on the Svetlana website and just stumbled across it. I've been curious about it ever since. I've seen mention of it a few times here since I joined this forum. But I haven't seen any build threads here yet. If you do build something with it, I'd be interested in seeing a thread about it.

Cheers,
Lou

[martin manning]

I think a 100W Fender Twin PT will work great for a pair of these, no issues that I can see. Each one kind of looks like two 6L6's. The only problem is that the plate load needs to be ~3-4k, so a a 100W Twin OT won't work. The quickest way around that would be to use a 100W Marshall OT and pretend it has 8, 16, 32 speaker taps to get 3k4.

Here are the curves...

[Andy Le Blanc]

yep... thats how I got the loading and power for KT90

I think I still have a tranny set for an old west 200w, the PT should be fine,
but the OPT is beast, way huge, impracticable for a combo, I know I have some twin iron hiding in a corner somewhere, the loading in the thing,
I'll have to play with numbers there and see what the best solution is on hand, tax time is coming so I can swing a OPT if I need to...

[martin manning]

Yes (duh), a Twin OT with double the speaker load would be good. Got a 4-ohm only Twin OT? At what voltage are you going to run the screens? Using the 250V curves with 425 or so on the plates and a 4k load looks pretty good.

[es345]

But I haven't seen any build threads here yet.

https://tubeamparchive.com/viewtopic.ph ... ight=kt120

[Andy Le Blanc]

well..., the kt120 screen rating is 600v, so with fender voltages it looks like I can start the screens off at plate voltage.
But if the build moves to higher plate voltages, the data sheet alludes to keeping the screens some 150v less than the plate (even at 400v plate).
My last experience with kt90 demonstrated that screens where happy around fender voltages, I suspect the kt120 will be the same.

I still have to do some digging to make sure it's not a bad assumption.

so far a 100w jcm 800 is the most affordable, and I've used it in the same application with kt 90.
1.7k primary that puts the load around 3.4 with a mismatch, a bit less probably with the actual voice coil, in the ball park any way. closer than a 2k primary fender opt.
The kt90 is constant use still some 5 years later, big fat clean thing...
It acts like a non-master twin, it breaks up very nicely but the volume, jeez...

[Andy Le Blanc]

Thats a good looking rig...

[Andy Le Blanc]

I've started to pa-ruse the hammond mfg PT lists and found a couple models that fit my old fart intentions.
the 372kx, and the 372lx... both are 600v ct, 50v bias, 6A 5v...
the kx has a 575ma rating and 9A fil., and the lx has 690ma rating and 12A fil.,..
Both will get me easily a max plate dissipation for two tubes without taxing the PT.
The voltage isn't imposing for DIY, the power rail should be easy and reliable.
The plate voltage meets my experiences with vintage gear thats been across my bench.

One question here that comes to my mind is if you can run two 5U4GB on 6A winding,
running two parallel drops the supply Impedance as well as increasing current capacity, I've been wanting to try it...

I've noticed in many old pieces that the PT was chosen despite the tubes.
The bias was for the power supply and not the tubes, the less taxed the PT the greater the reliability,
its a different ethic, I've got an old wards airline still running its 6l6g...

thinking out loud, I still would like to find a 3k opt that'll handle the current without mis matching

[es345]

Well, I am not shure that I have got all your assumptions, so I will just repeat, please correct if I am wrong. I have understood that you want to use

- a pair of KT120
- one of the 2 mentioned PT's
- ideally an opt with 3K Raa

I have made a short top down calculation based on the data of the second PT (600V AC, 690 ma). The excel is included. AC to DC is estimated 1,2 at full load based on some experience, efficiency of the power stage about 60%. Saturation voltage of the KT120 I have put in 60V based on the datasheets. That leads to a lower limit of Raa of about 3K5. If you go lower the PT will go into overload already at the max clean power.

If you then analyse the situation at the KT120 you will find:
- maximum plate dissipation will be violated, perhaps no so critical, you can do forced cooling
- more critical: at max clean power you reach the limit of the maximum average cathode current (case 1 clean). If you go for an overdrive (300%) you are going up to 340 mA, this shortly willl kill the tubes. (Case 1 overdrive)

So I would go for a opt of 5K Raa if you want to take KT120 and one of the mentioned PT's. Then you are on the safe side. (case 2)

[martin manning]

Andy, with the 600V CT are you planning to run FW rectifier and get ~400V B+, or FWB and get ~800V?

[Andy Le Blanc]

I think I see what You did...

your anode voltage...

600v ct, to a full wave capacitor input load, the same old PS,

The result is a peak voltage around .71 x the secondary, or 426v

so... 426 x .3 = 127 w for two tubes, and with a 60w anode and an 8w screen it actually falls short of max dissipation for the pair.

Why .3? I've found in many old things that have come across my bench for repair, a lovely coincidence between the operating point of the tube and the ma. rating of the PT.
With cathode bias, you bias hot and you mind where this is for your tube type in relation to the current load your putting on the PT.

I only want to bias around 50% of the transformer current rating, and have it correlate to a design center maximum dissipation figure.

I want the full wave capacitor input PS because it has 1:1 current relationship.

the tube run hot and the PT runs cool, tubes have sockets for reason.
They are disposable... the transformers are not.

proverb: chief old fart ranting


Thats a pretty slick software model, I'd love to see loading for 3k,4k, and 5k, Assume around 425 anode and screen V and a static bias .28A for the the pair.