Thanks Lars, - there are 50VA 115-0-115:6-0-6 for 17 euro each. These seem to be ideal with only a bit of unwinding possibly being necessary? I have Jordan JX92S speakers (8 ohm?) & Rogers LS3/5A (nominally 15 ohm) but I don't think this cahnges the require,ments or does it?
If you go for a 115V+115V:9V+9V you can create two taps for your different speakers.
Look into the TL783 for CCS duty. They are very easy to use and bias up consistently between items (unlike the mosfets). they have better specs than the LM317 and can handle 125V, and they are available from RS spares.They are more than adequate for this duty.
Shoog
Look into the TL783 for CCS duty. They are very easy to use and bias up consistently between items (unlike the mosfets). they have better specs than the LM317 and can handle 125V, and they are available from RS spares.They are more than adequate for this duty.
Shoog
Last edited:
Thanks Shoog, using the formula I still get 7V for 15 ohm speakers - what am I missing?
I'll leave the suitability of the TL783 for the more experienced to comment on.
Edit: Sorry mis-read my calcs - it comes out at 10V for 15ohm speakers - so maybe a 9-0-9 secondary would be about right. Can I unwind one side of a CT secondary?
I'll leave the suitability of the TL783 for the more experienced to comment on.
Edit: Sorry mis-read my calcs - it comes out at 10V for 15ohm speakers - so maybe a 9-0-9 secondary would be about right. Can I unwind one side of a CT secondary?
Last edited:
Which is why I would not completely count on the lower Crss specs. They may be better than the other part, may not.
Sheldon
jkeny,
Check this Nuvotem data sheet for a 50VA:
http://docs-europe.electrocomponents.com/webdocs/002e/0900766b8002e87b.pdf
To calculate impedance you use Noload Vsec. For parallell-connected secondaries using a 6V(6,8) and get 9,2k:8 and using series-connected you get 4,3k:15. A 9V(10,2) will give you 7,6k:15 with parallelled secondaries.
But.... I´d use the LS3/5A connected to a 8ohm output. It will sound thin connected to the 15ohm. Not the ideal speaker for a small tube amp.
Check this Nuvotem data sheet for a 50VA:
http://docs-europe.electrocomponents.com/webdocs/002e/0900766b8002e87b.pdf
To calculate impedance you use Noload Vsec. For parallell-connected secondaries using a 6V(6,8) and get 9,2k:8 and using series-connected you get 4,3k:15. A 9V(10,2) will give you 7,6k:15 with parallelled secondaries.
But.... I´d use the LS3/5A connected to a 8ohm output. It will sound thin connected to the 15ohm. Not the ideal speaker for a small tube amp.
Last edited:
I'll state this here again, because its critical information, the Telema's ring if the primaries are wired up in one direction, but don't if wired the other way. On the versions I have you need to wire them up like this;
Black - anode
White
+ - B+
Orange
Grey - anode
Unfortunately they aren't colour coded like that anymore, so if anyone has experience then chime in.
I just discovered that I had wired up one of my output transformers on my Tabor the wrong way round and as a result it was ringing. Changing it round produced a noticable smoothing of the sound.
Shoog
Black - anode
White
+ - B+
Orange
Grey - anode
Unfortunately they aren't colour coded like that anymore, so if anyone has experience then chime in.
I just discovered that I had wired up one of my output transformers on my Tabor the wrong way round and as a result it was ringing. Changing it round produced a noticable smoothing of the sound.
Shoog
Hey Shoog,
Really interesting info! But doesn´t this depend on which end of the secondaries you connect to ground?
Really interesting info! But doesn´t this depend on which end of the secondaries you connect to ground?
I have always connected my secondaries in parallel which means no to your question.
I think what we are dealing with is capacitance between the primary and the secondary. The primary is bifilar wound and built up from the core outwards, physically one end of the windings will be close to the core but distant from the secondary. Therefore one end will have significant capacitance to the secondary, and the other capacitance to the core but with nowhere to go so having minimal effect.
I performed a frequency sweep long ago and found the wrong way to peak out at about 50-60Khz and also to have a falling response below about 50hz. the other way round the response tails off from about 50kHz and is clean down to 10hz. I now choose by ear and the effect is quite noticable.
Shoog
I think what we are dealing with is capacitance between the primary and the secondary. The primary is bifilar wound and built up from the core outwards, physically one end of the windings will be close to the core but distant from the secondary. Therefore one end will have significant capacitance to the secondary, and the other capacitance to the core but with nowhere to go so having minimal effect.
I performed a frequency sweep long ago and found the wrong way to peak out at about 50-60Khz and also to have a falling response below about 50hz. the other way round the response tails off from about 50kHz and is clean down to 10hz. I now choose by ear and the effect is quite noticable.
Shoog
Last edited:
I'm just emerging from general anaesthetic! but I have been wondering why I only got about the same power as an SE amp in my build - EL86 210V/55mA 12-0-12 100VA toroid. As soon as I can lift my scope, I'll try increasing the current, and/or ordering another toroid. But 12V (2.9K Ohm) sounds about right for the EL86, unless I am missing something.
Any offers on how much power should be possible with EL86? will a 9V 30VA trafo work better (guitar amp will be fine to start with, so deep bass not critical for the next build)
It does not, as we normally assume parallelled windings with mains toroids as OPTs!
To clarify: With your secondaries in parallell have you checked if it matters which end of the parallelled windings you choose to ground?
Hey Rod,
Have you drawn your loadlines to check? For an SE EL86, 2,9-3k is OK just on the edge of crossing Pamax. Double it for PP Class-A. So 6k Raa should work better.
Speedy recovery, Rod - I hope al is well?
You've moved away from the ECL86 then?
Thanks Lars, Shoog,
Very useful information!
This is how it should be done using a descent triode driver(6C45 is also an option) together with an EL84. The alternative to a low Ri triode driver is a linear pentode like E280F, 6F12P etc.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Thanks - I should be back at the bench soon! A little surgery to fix a lump in the neck. A have a fine Frankenstein scar on the way, so I need to build more amplifiers with top-cap output valves - transmitting pentodes with 3kV crackling down the side to make the most of that 'monster lab' effect!
I still like the ECL86, but don't have a 6V trafo to hand - I thought my EL86 would drive the 12V toroid better, until I go shopping. Talking of which, what about these low cost parts?
http://www.rapidonline.com/Electrical-Power/Transformers/Toroidal/Toroidal-transformer/82719
Lars, thanks for the power sim!
Lars, Cool! 3 tubes for good stereo, low cost, current production valves
Is that 1V input for full power?
For pentode driver, we will need a source-follower to drive the Schade network, I assume
No I have not checked, but the secondary is so physically thin that I cannot imagine this is the source of the difference.
If my supposition is correct - it would be relatively easy to completely unwind the secondary and place a grounded screen between the primary and secondary. Personally I am happy with response out to 50K before roll of so I won't be trying the experiment.
Shoog
They look really good. There is of course the slight possibility that they wont perform as well as the Telemas, though this seems unlikely.
The only thing is, for my next project I need input transformers with a 2:1+1 ratio and the only ones in the range are 115+115:55+55 at 120VA. it seems a bit ridicules for the input transformers to be bigger than the outputs😀
Shoog
Clipping occurs with just above 1Vrms at the input.
On the contrary, we remove the 15k series resistor and let the pentode work like a currentsource driving into the, in this case ca 2k, load. But you need a pentode that is linear at a working point of 10-20mA. Should look something like this:
An externally hosted image should be here but it was not working when we last tested it.
Sensitivity will be very high at well below 0,2Vrms. Distortion will be low due to distortion cancellation.
Just add a CPU cooler with fan for the sand.....
Last edited:
- Status
- Not open for further replies.