| A choked Monkey, please! - Click HERE for Original Thread |
| Fuling |
Sorry for the very stupid subject title.
Less than two hours ago I finished my latest (prototype) amp which is a "monkey" topology using cheap Hammond chokes/transformers and a few not so extremely common triodes namely 6C4 and 12B4A.
To be honest the amp was finished yesterday but due to a bad PSU cap I didn´t get it to work until today.
This far I´ve only measured the output power (about 1,5W) but more measuring will be done this weekend.
Listening tests indicates the potential of something really really nice!
Bass response is a bit shy, but the imaging is next to holographic. |
|
|
| planet10 |
I found it kind of hard to read your digicam pic of your sketch so i took the liberty of cleaning it up -- if you could point out any glitches.
dave |
|
|
| Sch3mat1c |
OH NO!!!...6C4!!!...THAT'S A..12AU7!!!!!!! AAAAAARRRGGGGGGHHHHH!!!!!!!!!!!!!!!!!!!!!!!!!!!
lol.
12B4 is nice, except that it really sucks. :p It's plagued with low dissipation and mu, if you ask me. But if it weren't for that, I'd use it for everything. :)
Tim |
|
|
| EC8010 |
6C4 has no more right to be on God's clean earth than a weasel. It's half an ECC82 (which is a horrible valve).
Bung a 6J5GT in. Bodge a 6J5GT in. I don't care. You will see why 6C4 is best used as grinding paste.
12B4A is a bit different. Some are wonderful. Unfortunately, most aren't. You will probably want to select valves. |
|
|
| fdegrove |
Hi,
I don't care one bit about the u of the output stage, that should be dealt with in front of it.
The most linear devices inevitably have low u and low Ri, good thing too since you want current to drive the speakers, not voltage...would be useless in most cases anyway.
So, why not consider a 6C19Pi for output and a 6C5GT, for instance, at the input?
In the long run both types are readily available, not many people use the 6C5GT and AFAIK the 6C19Pi is still manufactured...
With CDPs at 2V RMS as a standard, how much amplification do you need from the amp?
Just my monkey,;) |
|
|
| Fuling |
Planet10: Excellent work and thank you, but there are some minor details.
The DC resistanses for the chokes are 595 and 3700 ohms. Can't blame you for reading the schematic wrong though!:cool:
I used these tubes because I already had them lying around (more or less at least). They look good together (wow, that´s one hell of an argument...) and I'm very pleased with the sonic results so I'm not going to change (unless someone makes me...)
Have a nice weekend, gents! |
|
|
| planet10 |
| quote: | Originally posted by EC8010
6C4 has no more right to be on God's clean earth than a weasel |
But you have to admire the ability to go to the pantry and whip something up that is satisfying with just what is there... ever had barbequed weasel? :)
| quote: | Originally posted by Fuling
The DC resistanses for the chokes are 595 and 3700 ohms. |
fixed
dave |
|
|
| Fuling |
I spent some time listening to this amp this morning, and I got a feeling that I'll never use plate resistors and coupling caps again:mafioso:
Right now I'm making plans for changing the driver stage in my 6B4G SE from 6SL7 (RC-coupled with cathode follower:dead: )
to a choke loaded 6SN7.
OK, there will still be a coupling cap but I see chances for great improvements. Changing it into a monkey would be too much work |
|
|
| nemestra |
| quote: | Originally posted by planet10
... ever had barbequed weasel? :)
dave |
Some friends of mine are just back from Peru and mentioned roasted guinea pigs ... One quick google later and
http://www.peru-travel-adventures.c...-11-5-2002.html
So Dave, what does barbequed weasel taste like?
James |
|
|
| planet10 |
| quote: | Originally posted by nemestra
So Dave, what does barbequed weasel taste like?
|
No idea.
But i do know that things we might consider strange to eat can be really good. The hamburger sundae for instance... sesame seed bun, slice of onion, hamburger pattie, a scoop of vanilla ice cream and chocolate sauce (Hershy's preferrably)
dave |
|
|
| EC8010 |
| No, no, what you want is a Camembert and banana sandwich (best in a French stick). |
|
|
| Brett |
| quote: | Originally posted by EC8010
6C4 has no more right to be on God's clean earth than a weasel. It's half an ECC82 (which is a horrible valve).
Bung a 6J5GT in. Bodge a 6J5GT in. I don't care. You will see why 6C4 is best used as grinding paste. |
Right on Brother EC.
But don't forget to always beware of the 12AX7, the devil-spawn of tubes. |
|
|
| fdegrove |
Hi,
| quote: | | But don't forget to always beware of the 12AX7, the devil-spawn of tubes. |
O.K., I agree as far as voltage amps go in amplifier and such but what else to use for a phonostage?
That is without introducing other malign artefacts such as microphony, noise and such...
We don't like the 12AX7,12AU7, the 6DJ8 is not trendy either....somehow, somewhere you need voltage amplification, right?
So what does get the green light?
Cheers,;) |
|
|
| Sch3mat1c |
This is where Choky, err I mean the ZM comes in and says something philosophical and Zen-like....
Tim |
|
|
| James D. |
Frank said...| quote: | | We don't like the 12AX7,12AU7, the 6DJ8 is not trendy either....somehow, somewhere you need voltage amplification, right? |
Well there is the E810F that Thorsten uses ...
or push-pull 6C45 give a stage gain of 100 and amazing transconductance and extra noise cancellation...
or 6AM6 10mA/V transconductance and mu of 90
or 6F12 (another special Russian) with a very high trans conductance and high mu triode and pentode in one tube...
I think there are lots of options for the DIY market. They do all have their pecularities but they all beat the pants off the 12AX7, 12AU7, 6DJ8 options...
ciao
James |
|
|
| EC8010 |
| quote: | Originally posted by James D.
...or push-pull 6C45 give a stage gain of 100 and amazing transconductance and extra noise cancellation... |
6C45 is a splendidly linear valve, but it still only has a gain of about 45, and a differential pair will always be 3dB noisier than a single-ended stage because the input is split between the two valves. Unless, of course, you were thinking of common-mode noise rather than thermal noise?
6DJ8 has a very nice combination of properties. It's not quite as low distortion as a 6J5GT, but it's not bad at all. Then there's 13D3, 37, and most of the triode-strapped pentodes. |
|
|
| Fuling |
Hey, stop bullying my 6C4´s
:bawling: :bawling: :bawling:
I must admit that I was completely wrong about the output power. I must have been a bit stressed when I measured it, cause I didn´t notice the difference between 1v/div and 2V/div on my scope...:dead:
Still, even though it puts out only 0,39W (with a far to weak PSU and very strange operating points for the tubes) this amp keeps me awake at night with an urge to hear more and more records.
Last night I found a new power transformer that will provide all the voltage and current needed with a bit of margin, so all I need now is some new PSU caps and a chassies. Currently everything is glued and screwed to a piece of chipboard... |
|
|
| James D. |
EC8010,
PP stage.
Common mode noise would cancel. But I was expecting thermal noise to cancel in the IT (after all it is random so half of it's one polarity whilst the other half is the opposite) to remove the 3dB disadvantage and leave a net decrease in noise... Another benefit of an IT stage that then drives the 600R LCR RIAA unit...:) :)
But I may be completely wrong here as I haven't modelled it or measured it... what do you think?
6J4, EC98, 6C37B would all work well as PP too and a whole raft of russian beauties...
the list goes on and on....
ciao
James |
|
|
| EC8010 |
Fuling,
if you can try a 6J5GT (or 7N7 or 6SN7 etc) in place of the 6C4, I'd be really interested to hear whether it makes your toes curl more, or less. The 6C4 could be cancelling distortion with the output stage, you see...
James,
sadly, no. Random noise is not coherent, so when you perform a subtraction (as in your inter-stage transformer) the noise powers add, lifting noise by 3dB. If you think about it, if both valves produce random noise, then their noise voltages at any instant must be different - if they were the same, or opposite, they wouldn't be random...
I've just had to look up EC98. Looks interesting. |
|
|
| James D. |
EC,
Now I need to go back to my text books...again:)
| quote: | | If you think about it, if both valves produce random noise, then their noise voltages at any instant must be different - if they were the same, or opposite, they wouldn't be random... |
Ok - but then how does the noise cancellation that happens when devices are paralleled into a load work...and with random noise on a signal isn't addition and subtraction the same for the noise from a cancellation viewpoint?
I never did understand noise cancellation theory...as you can tell:)
thanks
James |
|
|
| Fuling |
Well, I guess I could try a 6SN7 if I get time.
Would it be a VERY bad idea to split one between the channels?
I have done that with 6922s but they have screens between the sections, 6SN7s doesn't. |
|
|
| James D. |
| quote: | | Would it be a VERY bad idea to split one between the channels? |
No not at all. A little bit of crosstalk in a prototype to try something out is fine. Even in a finished unit the degree of crosstalk is relatively small. It's not best practice but it works fine.
As EC8010 said, it will be interesting to see if you think it is better or worse with the 6SN7....
ciao
James |
|
|
| Fuling |
Ok, I'll see if I can put one in there sometime before I build the finished version.
Though, I'm much more curious about how it would sound with the right points of operation. |
|
|
| EC8010 |
James,
when you add devices in parallel, it's easiest to get a conceptual handle on it by thinking about statistical surveys and samples. If you ask a single person a yes/no question, their answer is pure noise. When you ask a hundred people, any individual answer contributes a proportionately smaller amount of noise. Another way to look at it is to treat each device as a resistive noise source. When you add the resistances in parallel, the total resistance is much smaller, and it is perfectly valid to use this total resistance to predict the noise.
Fuling,
there's no problem at all in sharing a dual-triode between stereo channels. The capacitance between one anode and the other grid is so small that the crosstalk is tiny. People make a fuss about crosstalk, but LP, which has crosstalk of 35dB at very best, provides a perfectly acceptable stereo image. I'm almost tempted to set up a 6SN7 and measure the crosstalk... |
|
|
| Fuling |
I see.
I'm a bit tempted to share an 6922 between the channels for this amp.
The higher amplification factor and lower plate resistance would be of great use here I think.
The sad thing is that the 6C4s looks so damn good next to the 12B4s:mafioso: |
|
|
| Fuling |
I replaced those 6C4s with an PCC88 (poor mans 6922:devily: )
yesterday and...no, I preferred the 6C4s.
Strange, since the PCC88 seems to suite much better for the application when looking at the specs. |
|
|
| dhaen |
Hi,
It can happen.:nod:
Did you get the heaters hot enough?
BTW Your original schematic looks as if it's drawn on a fingertip, and zoomed in...
Cheers, |
|
|
| Fuling |
About the heaters: Yes I hope so:)
About the schematic: In complete lack of drawing software I drew it up on a piece of paper and toock a photo with my webcam, thereof the extremly bad quality!
;) |
|
|
| dhaen |
Hi,
What I meant was that the PCC88 has 7.6 volt heaters.
The drawing was just fine...:rolleyes:
Cheers, |
|
|
| Fuling |
Some datasheets says 7,6V, some say 7V, I've always used them with 6,3V without problems.
It's not an issue after all, since I'm not going to use it anyways.
I´ll probably stick to the 6C4´s since I really liked them and the heater winding in my power trannie is already running on the edge, so a 6SN7 would broably be too much for the poor bastard.
Now I expect someone to say "Use an additional heater transformer then!!!"
No I won´t, there´s already too many inductive components in this amp that will have to be placed in different angles and at adequate distance from each other.
I´ll save my NOS 6SN7s for my 6B4G amp instead;) |
|
|
| fdegrove |
Hi,
| quote: | | Some datasheets says 7,6V, some say 7V, I've always used them with 6,3V without problems. |
Shortchanging the heaters on voltage or current won't prevent the valve from working but...just don't expect anything close to good sound from that arrangement.
Just that you know,;) |
|
|
| Fuling |
Well, another reason to use 6C4´s then!:devily:
/Zoega, för hälsan och avföringen! |
|
|
| planet10 |
Fuling,
Could i get you to turn off the "smart" quotes in whatever you are using to edit messages -- the Yen signs are starting to drive me bonkers :)
(typesetter's quote marks are a non-standard ASCII and differ from platform to platform -- they probably look like quotes to you)
dave |
|
|
| Sch3mat1c |
Works for me.
C'mon...get IE... it won't kill ya......trust meeee....... :devilr: :goodbad: :p
Tim |
|
|
| Fuling |
Hmm, that didn´t look as it did when I wrote it...
Well well, I´ll quit using the ´:)
IE??? |
|
|
| fdegrove |
Hi,
| quote: | | Hmm, that didn´t look as it did when I wrote it... |
Well, I don't see that yen sign in your posts either.
IE = Internet Explorer, by Mickeysoft.
If your running any version of Windozs you can download it for free at:
MICROSOFT.
Cheers,;)
EDIT:I don't see any reason the internet browser soft is at fault here, it may just as well be something totally different.
Maybe Dave sees a pattern? |
|
|
| planet10 |
| quote: | Originally posted by fdegrove
EDIT:I don't see any reason the internet browser soft is at fault here, it may just as well be something totally different. |
it isn't the browser -- it is the platform... windows vrs the rest of the world. If i type a curly quote i'll see it fine, you guys should see something else
Õ <-- curly quote
There are also lots of other useful characters that are not usable.
¡ <-- degrees
µ <-- mu
and when PC guys use the built-in symbol for fractions it also comes out goobledegook.
dave
planet10/very happilly running the most common UNIX OS.
(halo we miss you) |
|
|
| fdegrove |
Hi,
Dave:
As you may recall I run Mickeysoft Win2K.
So far anything you typed seems fine to me and I assume it's like that to other Windozs users.
But:
This is actually a capital O with a tilde on top.
No, a plain i.
Yes that is the greek alphabet mu.
Now if I retype what Fulling wrote:
6C4' s then logically you'd see a Yen sign where the space is.
All this assuming Fulling is running Windozs, I'd like to know what version since this may be language (version) related and a patch may be available....
Cheers,;) |
|
|
| planet10 |
| quote: | Originally posted by fdegrove
Yes that is the greek alphabet mu. |
Surprise!! a coorespondance.
All ASCIIs below 127 are standardized (7 bit), all above (8-bit) are platform specific, Apple started using them 1st *(and then M$ choose an incompatible set), hopefully things will smooth out again when unicode becomes the practical standard...
dave
*at least amounst the more widespread vendors. There are still some routers on the internet which can't pass 8-bit which causes some havoc. |
|
|
|
|
