Here goes...
and ignore the feedback cap, its should be 270pf ish - and error before i uploaded.
Second error - the output css will of course be bypassed.
![IMGDEAD]](/community/proxy.php?image=http%3A%2F%2F%5BIMGDEAD%5Dhttp%3A%2F%2Fimg39.imageshack.us%2Fimg39%2F9166%2Fel34ampv3.jpg%5B%2FIMGDEAD%5D&hash=b9a3232adcc4f3cbaba069752cb5036d)
and ignore the feedback cap, its should be 270pf ish - and error before i uploaded.
Second error - the output css will of course be bypassed.
Last edited:
Although I think that using a CCS in a push-pull output stage is not a good idea in general (others disagree), separate ones like that will cause the stage not to work.
One more bit of lily-gilding- you can replace the RC in the cathode of the input tube with an LED.
edit: saw your edit about the CCS.
One more bit of lily-gilding- you can replace the RC in the cathode of the input tube with an LED.
edit: saw your edit about the CCS.
how does led biasing affect the feedback network
edit - thinking about it i just need to move the feedback connection directly to the cathode.
edit - thinking about it i just need to move the feedback connection directly to the cathode.
Last edited:
If you put back to back caps (something like 500-1000uf each) between the cathodes and place a bleeder resistor at the junction (something between 300R and 1Meg) you will have an output stage which behaves somewhere between a standard output and a differential stage. The caps can be relatively low voltage if you use a 1meg resistor - think 100V.
Shoog
Shoog
getting closer....
An externally hosted image should be here but it was not working when we last tested it.
Two steps forward, one back. You still need the bottom resistor of the feedback divider- the LED has close to zero AC impedance (which is why it substitutes for the RC combo).
yes - just worked that out before you replied.
I think... give or take a view rather important values i am about there. I'll post up a schematic of the ps and CSS tomorrow and see if i have done anything wrong.
This is a great learning experience, so thanks for helping out.
I think... give or take a view rather important values i am about there. I'll post up a schematic of the ps and CSS tomorrow and see if i have done anything wrong.
This is a great learning experience, so thanks for helping out.
An externally hosted image should be here but it was not working when we last tested it.
i would use r-c type bias to start with (but i am a sad old far.....}
Mullard 5-20 Audio Hi-Fi Amplifier
up until the mid 1950s mullard used resistor 5 on there scemo {use link }
to set their amps to the speaker impedance.
but you should be able still to use c9-r13 type set to do this
as mullard did from the 1950s onwards {as in link }
Mullard 5-20 Audio Hi-Fi Amplifier
up until the mid 1950s mullard used resistor 5 on there scemo {use link }
to set their amps to the speaker impedance.
but you should be able still to use c9-r13 type set to do this
as mullard did from the 1950s onwards {as in link }
pointy, I'll be trying both. I tried both with the baby huey (ginger tubes design) and the css sounded significantly (read noticeably) better.
But, both can be tried.
But, both can be tried.
indeed.
The 6n1p was chosen because i have a few, and its cheap. However.... they want a higher plate and more current than i am giving them here.
The input stage will probably have to be a 6dj8 in reality. 140v plate, 2-5ma, 2v grid.
The splitter.... the css helps a lot, maybe i can get away with the values given?
what do you think?
The 6n1p was chosen because i have a few, and its cheap. However.... they want a higher plate and more current than i am giving them here.
The input stage will probably have to be a 6dj8 in reality. 140v plate, 2-5ma, 2v grid.
The splitter.... the css helps a lot, maybe i can get away with the values given?
what do you think?
Css's. These may be wrong, particularly the cascode
for the LTP
For the output bias
Will these work?
for the LTP
An externally hosted image should be here but it was not working when we last tested it.
For the output bias
An externally hosted image should be here but it was not working when we last tested it.
Will these work?
The cascode you have should give you about 5mA. Lovely! For the output stage, you can use the same topology but with higher current devices. They can be slow and cheap if you're going to bypass, though why this is better than the classic automatic bias (a plain old bypassed resistor) still escapes me. One other fun thing- you can replace the base-to-base resistor with a green LED. It probably won't work any better for that application, but I love having more glowing bits.
For the input stage, the ECC88/6DJ8 works OK, but needs more current and perhaps less plate voltage. 6-8mA and 100V is a reasonable spot. Other good candidates there would be 6SL7, 12AT7/ECC81, triode-connected EF86, or if you really want to shoot the moon and see how good you are at keeping things stable, D3a. No doubt there's a zillion others (and I admit my unfamiliarity with most Russian tubes). For both the input and the splitter, you'll want to get out the plate curves and a ruler and start playing with load lines.
For the input stage, the ECC88/6DJ8 works OK, but needs more current and perhaps less plate voltage. 6-8mA and 100V is a reasonable spot. Other good candidates there would be 6SL7, 12AT7/ECC81, triode-connected EF86, or if you really want to shoot the moon and see how good you are at keeping things stable, D3a. No doubt there's a zillion others (and I admit my unfamiliarity with most Russian tubes). For both the input and the splitter, you'll want to get out the plate curves and a ruler and start playing with load lines.
"playing with the ruler"
if V1 is a 6dj8,
HT= 320
RL= 33k
ledbias -2v
gives about 10ma and looks ok.
The 6n1p, 8ma (css) looks ok with 175 plate. could be better, trouble is that would mean spending!
if V1 is a 6dj8,
HT= 320
RL= 33k
ledbias -2v
gives about 10ma and looks ok.
The 6n1p, 8ma (css) looks ok with 175 plate. could be better, trouble is that would mean spending!
can anyone recommend a different tube to the 6n1p for the splitter, i havent got enough HT headroom to get a good operating point. I dont want to use 6dj8's throughout as they cost too much.
must be a small 9 pin (b9a) with 6v heater
needs to be happy at 130 - 150 plate v
Css in the tail.
edit... 12at7 looks like a potential candidate (6.3 series)
must be a small 9 pin (b9a) with 6v heater
needs to be happy at 130 - 150 plate v
Css in the tail.
edit... 12at7 looks like a potential candidate (6.3 series)
Last edited:
???
6n23p, great deals on Electronics, Computers Networking on eBay!
Looks cheap enough to me😉.
http://klausmobile.narod.ru/testerfiles/6n23p.htm
Last edited:
6DJ8s are not expensive unless you insist on collectors' items. 12AT7 is a fine tube for the splitter- run it at 2mA, 150V on the plate, plate load at >100k, and you'll get a very favorable distortion spectrum.
i am dc coupling from the 6d78 stage, so the grid will be at 100v, so i need 250v on the plate. Ra will be droping 139v. @2ma the plate load becomes 69k. is that not enough. OR is my maths hopeless.
Ideally, you'd want the plate load resistor to be higher than that, which unfortunately means a higher supply voltage (450V with a 100k load) to that stage. Plate loading with a CCS will drop that requirement at the expense of complication... The other possibility is dropping the voltage on the plate of the 6DJ8- it will be fine at 70V, which gives you a little more room to play with.
- Status
- Not open for further replies.