Over the past year, I have created 2 different pairs of monoblocks:
1. Lorzenz & Siemens C3g driving EML45B
2. Valvo Aa driving EML50
The C3g/45B amp was good, but lacked the transient snap and "liveness" that gives jump to the music. I made a few changes to the circuit including upgrades to power supply caps & DC filaments. These were nice upgrades, but still no nirvana.
So, I had a pair of Valvo Aa and EML50 tubes and put them to work. The Aa has a mu=30, but the Rp = 30K. So, the interstage transformer (5K, 28ma) that was used in the C3g/45B amps would not be a great match for the Aa driver. The IT was replaced by a 0.22uF / 250K RC coupling network. Note: the EML50 can be used with a maximum value = 250K grid leak resistor. How does the Aa-->EML50 sound? Fair at best. The Aa has better transparency, harmonics, tonality than the C3g. But, the sound is rather flat. Not sure if this is due to the RC network instead of the IT. Actually, I tried the Aa --> IT --> EML50 and this IT / Aa mismatch sounded better than the RC coupling. Go figure.
Anyways, as much as I would like to use the Aa, the high Rp is becoming a problem with an IT. And, I don't like RC networks and am trying to minimize the use of caps and resistors in the signal path.
So, any thoughts on replacing the RC coupling between the Aa and EML50? Otherwise, I need to consider ditching the Aa and find a new driver tube. Anyways, searching for ideas.
Current schematic is attached.
1. Lorzenz & Siemens C3g driving EML45B
2. Valvo Aa driving EML50
The C3g/45B amp was good, but lacked the transient snap and "liveness" that gives jump to the music. I made a few changes to the circuit including upgrades to power supply caps & DC filaments. These were nice upgrades, but still no nirvana.
So, I had a pair of Valvo Aa and EML50 tubes and put them to work. The Aa has a mu=30, but the Rp = 30K. So, the interstage transformer (5K, 28ma) that was used in the C3g/45B amps would not be a great match for the Aa driver. The IT was replaced by a 0.22uF / 250K RC coupling network. Note: the EML50 can be used with a maximum value = 250K grid leak resistor. How does the Aa-->EML50 sound? Fair at best. The Aa has better transparency, harmonics, tonality than the C3g. But, the sound is rather flat. Not sure if this is due to the RC network instead of the IT. Actually, I tried the Aa --> IT --> EML50 and this IT / Aa mismatch sounded better than the RC coupling. Go figure.
Anyways, as much as I would like to use the Aa, the high Rp is becoming a problem with an IT. And, I don't like RC networks and am trying to minimize the use of caps and resistors in the signal path.
So, any thoughts on replacing the RC coupling between the Aa and EML50? Otherwise, I need to consider ditching the Aa and find a new driver tube. Anyways, searching for ideas.
Current schematic is attached.
Attachments
As euro21 says, consider Ale's gyrator. That would work.
I have only a basic knowledge of solid state so I tend to use plate chokes and I've been very happy with that solution. My plate chokes all have 150H inductance or more and I should think the Aa would be happy with that kind of load. Consider the Lundahl LL1668 (not LL1667 which has high resistance). Or why not try out the Hammond 193A as a cheaper solution - that probably would sound great - 200H. Buy a pair and tell us how it works - you might start a trend!
193A - Hammond Mfg.
I have only a basic knowledge of solid state so I tend to use plate chokes and I've been very happy with that solution. My plate chokes all have 150H inductance or more and I should think the Aa would be happy with that kind of load. Consider the Lundahl LL1668 (not LL1667 which has high resistance). Or why not try out the Hammond 193A as a cheaper solution - that probably would sound great - 200H. Buy a pair and tell us how it works - you might start a trend!
193A - Hammond Mfg.
Andy, for Ri:30k at least 500H inductance fulfils the -minimum:2x Ri at 20Hz- requirement of load.
With this high output impedance tube "mission impossible" to use inductive load.
With this high output impedance tube "mission impossible" to use inductive load.
Thanks for that - I didn't do the maths and I haven't tried anything over 12K myself.
But do we really need to plan for 20Hz? Low E on a bass is 41Hz. I'm a bass player so I know these things! Even assuming low B which is the lowest you're likely to hear on a 5 string, then that's 31Hz. OK - humans can potentially hear down to 20Hz but there's no relevant fundamental notes under 31hZ, and mostly 40Hz. And anyway the harmonics will fill in the sound for the odd note on the grand piano. Plus how many speaker systems are flat to 20Hz?
So how does the maths look at 30Hz and at 40Hz?
Z= 2*Pi*f*L
If Z=60k
at 30Hz 318H
at 40Hz 239H
requiring.
But large inductance has to many turns (with enough large capacitance), so current via these inductors generally limited.
BTW if you locate -3dB point upper than "ordinary", the phase also begins to deviate upper, so begins to smearing the lower octaves.
No.
Uamax is 250V.
Use it at 160V, -2V as Ale suggested (minimum distortion).
The power supply voltage of this stage would be
2V+160V+(Ia*RDC of choke).
BTW without grid leak resistor the 50 tube will be destroying.
Beware, that you don't use NOS #50 tube with such high grid leak resistor as EML suggest!
p.s.
read this topic:
Type 50 Tube Amp
Uamax is 250V.
Use it at 160V, -2V as Ale suggested (minimum distortion).
The power supply voltage of this stage would be
2V+160V+(Ia*RDC of choke).
BTW without grid leak resistor the 50 tube will be destroying.
Beware, that you don't use NOS #50 tube with such high grid leak resistor as EML suggest!
p.s.
read this topic:
Type 50 Tube Amp
Last edited:
Here's a slightly corrected diagram, just some basics. If the 50 is in self bias, then the bias is 81v on the cathode, not -81v. Same for the input tube, plus 2v. You need a cathode bypass cap of at least 40uF on the 50, not 5.6uF. I strongly suggest a DC Link cap here.
You can work out the dropper resistor to the plate choke with Ohms Law. First you need to determine the voltage drop through the plate choke.
You can work out the dropper resistor to the plate choke with Ohms Law. First you need to determine the voltage drop through the plate choke.
Attachments
Thanks Andy. The EML50 (in auto bias) has a max allowable 250K grid resistor. That said, the Rp of Aa = 30K. So, 30K x 3 = 90K. I could go with a 100K grid resistor.
I will change the cathode bypass to 50uF WIMA DC Link.
Q: should the B+ for the plate choke come directly from the LCLC instead of thru the output transformer?
The problem that I have with the schematic is that the signal is still going thru an RC network, which is what I was trying to avoid. I am trying to minimize Caps and Resistors in the signal path. I may have to rethink the Aa --> EML50 and look into a different solution. Damn, can't have everything!
I have seen a nice Aa --> GM70 design (direct coupled) that could be pursued. This would require a new power supply to support a 1KV B+, but it might be rationale for getting new chokes and power transformers from Alexander. Very tempting. The Aa tube is really nice, thus my desire to use it. Just thinking out loud.
I will change the cathode bypass to 50uF WIMA DC Link.
Q: should the B+ for the plate choke come directly from the LCLC instead of thru the output transformer?
The problem that I have with the schematic is that the signal is still going thru an RC network, which is what I was trying to avoid. I am trying to minimize Caps and Resistors in the signal path. I may have to rethink the Aa --> EML50 and look into a different solution. Damn, can't have everything!
I have seen a nice Aa --> GM70 design (direct coupled) that could be pursued. This would require a new power supply to support a 1KV B+, but it might be rationale for getting new chokes and power transformers from Alexander. Very tempting. The Aa tube is really nice, thus my desire to use it. Just thinking out loud.
Look, I just did a quick diagram to try and help out, nothing more. So please elaborate and post a diagram of what you see as the correct circuit. Thanks!
The dropper resistor (marked "?") in the schematic in post #12 should be connected to B+, so not to the plate of the EML50.
Ah yes. Thanks. I always draw the output transformer the other way up with the B+ on the top like most schematics so I didn't spot that in my haste.
Attachments
Last edited:
Taking driver B+ from the plate of the power tube is not wrong. It is just a different solution. It gives less power supply ripple, but it gives some negativ feedback. Most of it disappears with the decoupling cap though. But the gain og the Aa tube might be a tad low for this solution. 🙂🙂
"It gives less power supply ripple, but it gives some negativ feedback. "
Are you kidding? 😛pp
There are few hundred Volt peek_to_peek signal.
Negative feedback as poor PSRR of driver stage ... it's a joke. 🙂
Are you kidding? 😛pp
There are few hundred Volt peek_to_peek signal.
Negative feedback as poor PSRR of driver stage ... it's a joke. 🙂