Hello everyone,
I tried to design an output stage for my TDA1541 dac using a D3a. Well, not really designing this time (my second), I got most of it from the data sheet..
But I couldn't completely copy it, as some values aren't mentioned..
The troublemaker here is R6, I haven't got a clue how to determine this one, couldn't find it in my books..
Also I chose 1K for the gridstopper R2, but it's kind of a guess..
Designing this stuff is quite new to me, I'm terrible at maths but I'm eager to learn. I'm sure posting this here will get me some usefull extra info. All comments are very welcome!
Thanks in advance!
I tried to design an output stage for my TDA1541 dac using a D3a. Well, not really designing this time (my second), I got most of it from the data sheet..
But I couldn't completely copy it, as some values aren't mentioned..
The troublemaker here is R6, I haven't got a clue how to determine this one, couldn't find it in my books..
Also I chose 1K for the gridstopper R2, but it's kind of a guess..
Designing this stuff is quite new to me, I'm terrible at maths but I'm eager to learn. I'm sure posting this here will get me some usefull extra info. All comments are very welcome!
Thanks in advance!
R6 is usually defined by the tube (next stage resistor), but it is not critical. The main function instead of defining output impedance, also to discharge the coupling capacitor if it is disconnected from next stage.
Usually, a value around 100k-470k is enough.
Grid stopper is commonly using 1k, carbon composite (like AB, Riken, Takman TRex), or Tantalum (AN, Sinkhoh). Can be ranged from about 200 ohm or up. Lower value may increase a possibility to self-oscilate(unstable). But the higher, may make high freq roll off.
Another comment, Input Coupling 100nF may be to small (to high of low freq roll off), as the grid leak is about 47k. I like about 1uF (PIO or good MKP), or bypass it as long as there is no DC offset from previous stage.
Thanks,
Ervin L
BTW, I'll be using a very low value I/V-resistor, I'm not quite sure if the operating point is best suited for this. I was looking at the curves, and liked the -0.5V line at around 95V and 30mA.. In the end I decided the TFK guys probably knew their stuff a lot better than I do.
http://www.hifitubes.nl/weblog/wp-content/telefunken-d3a.pdf
http://www.hifitubes.nl/weblog/wp-content/telefunken-d3a.pdf
I've done some experiments before, started with 100 ohm as suggested on Lampi's website, went lower step by step and settled with 28 ohm if I remember correctly. 100 ohm is waàày too high btw, you get a nice distortion 'sauce', and a decent output voltage, but it's very much 'my-fi'..
I believe some people even used 18 ohm..
I believe some people even used 18 ohm..
Isn't the plate resistor way too low in value? If you don't have more B+ for the stage I would recommend a CCS or gyrator instead of that 625R thing (which is about 1/4 of the D3a internals resistance, when triode connected).
Thanks, this is the kind of comment I've been aiming for 🙂
And I'm still in the dark about what a plate resistor actually 'does', apparantly it can have a 'too low' value. What consequences would this have?
I've got my eyes on a small 125 V tranny, it would fit in my cd player, that's why.
Making the plate load too small increases distortion, and compromises output swing. You don't want a steep loadline like that. You may not get the output you need, and it'll sound like
An externally hosted image should be here but it was not working when we last tested it.
In this case, I'd go with an active plate load. That will maximize output swing, and minimize distortion.
Well, the higher the value of the plate resistor, the higher the gain (up to the mu of the valve employed), the lower the distortion and the higher the input capacitance (due to miller effect).
I recommend you to have a read on chapter 2 of the 3rd edition of "valve amplifiers" by Morgan Jones. It explains loadlines very very well! After reading a couple of pages you can print the triode curves of the D3a and start 'old school' electronics with a ruler and a pencil!
I recommend you to have a read on chapter 2 of the 3rd edition of "valve amplifiers" by Morgan Jones. It explains loadlines very very well! After reading a couple of pages you can print the triode curves of the D3a and start 'old school' electronics with a ruler and a pencil!
Ah yes, of course, the value determines the angle of the load line..
I studied Jones for the preamp I made yesterday, I'll get back to it.
I chose an operating point at first for the D3a, -0.5V 30 mA at 95V, but then I got afraid it was too weird and went with the Telefunken advice.. The voltage swing at the input is very small (unfortunately I can't measure that yet) so I guessed -0.5V was okay.. And my instinct tells me it's always beneficial to pump a lot of milliamps..
C3 and R6 form a highpass circuit. If C3 is fixed, R6 determines how much low end is cut.
For example if R6 is 1k, the -3dB frequency will be 338.63 Hz.
If R6 is 22k, -3dB frequency will be 15.39 Hz.
If you want a low value for R6 with good bandwidth, C3 needs to be bigger; for example
C3 = 10uF
R6 = 1k
gives -3dB at 15.92 Hz. Of course that low R6 wouldn't be practical for other reasons.
For example if R6 is 1k, the -3dB frequency will be 338.63 Hz.
If R6 is 22k, -3dB frequency will be 15.39 Hz.
If you want a low value for R6 with good bandwidth, C3 needs to be bigger; for example
C3 = 10uF
R6 = 1k
gives -3dB at 15.92 Hz. Of course that low R6 wouldn't be practical for other reasons.
- Status
- Not open for further replies.