Adjusting LTP values (12AT7) to account for lower B+ values

[BrownIsound]

Hi,

I have made a home brew amp using 2 6BM8 tubes, assuming the triode halves act like a 12AT7, driving the pentode halves as a push pull pair, cathode bias. I used the silverface LTP values for the PI (47K plate, 270/22K tail), but I wonder if since I have quite a bit lower B+ than a silverface fender (about 210V B+ for the PI node) if the values could be adjusted to get better drive for the power tubes, accounting for the lower plate voltage.

Is there any rule of thumb for voltage/current for optimum LTP function, or values should be adjusted x amount for y change in B+ voltage?

Sorry if this is something that has been rehashed a lot, I did a search and read up on LTPs, but couldn’t really get the answer to my specific question (or maybe I’m too dense to get the answer from the other threads). Also, is the assumption that the triode half of the 6BM8 acts the same as a half a 12AT7 correct?

[Stevem]

You might be interested in looking at the schematic for a Univox U45B model which can be found on line.

This uses the triode sections as a driver and a splitter.
It runs on a little under 300 volts to the output tube plates.

I have two of these amps and every one who plays it cranked is amazed at how it sounds like a little Marshall and cleans up great with the volume on the guitar .

[wpaulvogel]

You should use 100k plate resistors for more gain. I’d also use a 470 ohm cathode bias resistor and 10k tail.

[roberto]

You can use a CCS instead of the tail resistor, and get back some voltage across the triodes (plate is already lower, this way you lower cathode voltage as well).
If you do so, use the same vale for both the plate resistors of the LTP.

[BrownIsound]

Trying to keep it as simple as possible, not much room under the hood. Will likely try experimenting with the resistor values on the PI for now.

On a similar note, a post I found had an interesting point about connecting the cathode of the PI to a negative voltage (bias supply) to get the current/voltages more in line with a higher B+ amp. I wonder if you could connect this lower potential to all the preamp tubes to have the preamp/PI more like a “normal “ high power amp, connected to the lower voltage/lower power tubes.

Any noise concerns with this arrangement, especially concerning the first preamp tube?

[Ten Over]

Will likely try experimenting with the resistor values on the PI for now.

Try what wpaulvogel suggested. Those are exactly what I would have suggested had I been inclined to post (which I usually am not).

[BrownIsound]

Yes, I was planning on starting with his suggestions. Haven’t had time to mess with it yet, hopefully sometime this week.

[roberto]

Standard Marshall's 12AX7 LTP values are far from optimal for the curves of a 12AT7, on top of that the tail value is too low to have 100k on both plates and keep the balance.

[martin manning]

Here are a couple of on-line design guides for LTP:

http://www.valvewizard.co.uk/acltp.html
https://www.aikenamps.com/index.php/des ... e-approach

[martin manning]

Below is an example of a LTP with a simple constant current source to maximize Va-k for a given HT supply voltage (see at the bottom of this page: http://valvewizard.co.uk/ccs.html).
You can use this circuit as shown with a 12AT7, but for your 210V PI node voltage, make the plate loads 75k. R1 can be 100k 1/2W, and a common 2N3904 NPN transistor will be fine. If you have NFB, you can experiment with putting that in at the top of the emitter resistor, which can be 680Ω if that's what you have. You can use a red LED in place of the three 1N4148 diodes to keep the part count down.

[roberto]

Even better if you:
- add a 500 Ohm pot with the wiper on the CCS and the two side on the cathodes of the 12AT7 to balance them;
- use a cascoded CCS;
- use a red led instead of the multiple diodes and a red led inbetween the two transistors.

I'd also stay with 47 to maximum 68k on the plates, and increase the CCS current instead.

[BrownIsound]

Here are a couple of on-line design guides for LTP:

http://www.valvewizard.co.uk/acltp.html
https://www.aikenamps.com/index.php/des ... e-approach

I did the calculations based on the valvewizard link to get a few scenarios. I’ll try some of them out this weekend, if the yard work doesn’t consume all my time.

[BrownIsound]

I’m not against doing a CCS, but since I have resistors on hand, that’s the easy way for now.

I have only tried one of the calculated scenarios, but so far like it a lot, so I’ll play it during practice tomorrow to get a good feel of it before deciding to change it again.


Ended up with 700/10k for the bias/tail, and 60K plate loads.
I also have negative feedback injected at the tail with a 55ohm resistor to ground that is bypassed with a cap (can’t remember it’s value or the resistor from the ot at the moment).

Much better bandwidth and less compression than what was there before. I have a ppiv mv so it is easy to hear how the pi distorts without cranking up the amp.

I had originally put in the bypass cap on the feedback to give it more presence (like a Marshall presence on 10) but I think I may take it out now, as the new values definitely increased the brightness.

My first impression when I played it with the new values was “this sounds like a BF/SF Fender now”. Ha ha.