higher currents
Hi,
I have tried currents up to a bit over 90mA. That is all I can take out of my variable power supply. I really need to get to work breadboarding the power supply with my real transformer, etc. so I can verify what sort of supply I will have available. For the time being I have a 6kV PIV 0.6A bridge that should work, shouldn't it?
Although odd order harmonics decrease somewhat with increased plate current, the effect does not seem to be substantial in the area of positive grid current. However, if run at low currents, ie <70mA/tube higher order distortion increases significantly. I also would like to try the 1620 at 6.6k output impedance, just to see what it measures and listens like. Kind of a pain to rewire else I would have done so already. It does seem that the bass may decrease some at higher B+. Any ideas?
Michael
Hi,
I have tried currents up to a bit over 90mA. That is all I can take out of my variable power supply. I really need to get to work breadboarding the power supply with my real transformer, etc. so I can verify what sort of supply I will have available. For the time being I have a 6kV PIV 0.6A bridge that should work, shouldn't it?
Although odd order harmonics decrease somewhat with increased plate current, the effect does not seem to be substantial in the area of positive grid current. However, if run at low currents, ie <70mA/tube higher order distortion increases significantly. I also would like to try the 1620 at 6.6k output impedance, just to see what it measures and listens like. Kind of a pain to rewire else I would have done so already. It does seem that the bass may decrease some at higher B+. Any ideas?
Michael
one other thing
While the performance of the 12B4 driver was good with a 10k load, it was much better with a 60k load at large rms outputs at least. Maybe I should try a different driver. I have a pair of 6B4G I could probably wire up in an hour or two. What do you think? 250V plate 40-50mA fixed bias?
Michael
While the performance of the 12B4 driver was good with a 10k load, it was much better with a 60k load at large rms outputs at least. Maybe I should try a different driver. I have a pair of 6B4G I could probably wire up in an hour or two. What do you think? 250V plate 40-50mA fixed bias?
Michael
ciao Michael
I did not find plate curves of the 6b4g but if it is
similar ( I read so) to a 2A3, then it is, by far, a better tube
with respect to 12b4a.
Direct heated triode are more linear, as you know.
I will try some simulation with 2A3.
bye
Federico
6B4G
Hi Federico,
Yes, it is the same as a 2A3. If distortion is a measure of non-linearity, it is not too bad for the 12B4. Since I have the tubes it would be fun to try them. They do add a fair amount of expense and bother to the design though.
I expect they will model well, but then so did the 12B4. I still wonder if I don't somehow have something wrong. This even though the circuit sounds wonderful.
Michael
Hi Federico,
Yes, it is the same as a 2A3. If distortion is a measure of non-linearity, it is not too bad for the 12B4. Since I have the tubes it would be fun to try them. They do add a fair amount of expense and bother to the design though.
I expect they will model well, but then so did the 12B4. I still wonder if I don't somehow have something wrong. This even though the circuit sounds wonderful.
Michael
6B4G
Hi Federico,
Any chance to model the 6B4G yet? I wonder how the operating point will work?
I have been ill for the past couple of days so I haven't had a chance to try it yet, but hopefully today or for sure this weekend.
Michael
Hi Federico,
Any chance to model the 6B4G yet? I wonder how the operating point will work?
I have been ill for the past couple of days so I haven't had a chance to try it yet, but hopefully today or for sure this weekend.
Michael
Hi Michael
I have been ill too, and I'll be very busy in the next few days.
I hope to have the 2a3 modelled for the end of the week.
cheers
Federico
I have been ill too, and I'll be very busy in the next few days.
I hope to have the 2a3 modelled for the end of the week.
cheers
Federico
Hi
However, consider that 2a3 has a lower gain, about 1.5 times lower than that of the 12b4a.
I know that a good operating point for 2a3 when used
as an output tube is Vak about 310-320, bias -60, current
45mA. I do not know if it is good also when it is used has driver.
When I'll have the model I'll make some trial.
Federico
However, consider that 2a3 has a lower gain, about 1.5 times lower than that of the 12b4a.
I know that a good operating point for 2a3 when used
as an output tube is Vak about 310-320, bias -60, current
45mA. I do not know if it is good also when it is used has driver.
When I'll have the model I'll make some trial.
Federico
2A3
Hi,
Unfortunately, I didn't get anything done this weekend, so I haven't tested the 6B4G yet. I am concerned about the gain loss as well. Actually, I have been thinking about that since last week when I was listening to some classical piano that was recorded at quite a low level. At max input I could still only get moderate volume out. As I recall from some testing, about 1.1Vrms input drove the 845 to the limit of A1 when it was biased at -125V. I thought that was sufficient gain, as 2Vrms would probably then drive it to the limit of A2, if I could get that working. I was generally trying with this design to have the gain matched to CD player output levels so that minimum attenuation could be used. But that low output level CD kind of upset that thinking. One thing I could do, I think this works, is put a DPDT switch into the secondary side of the LL1676 that would change the input transformer from 1:1+1 to 1:2+2. I have read a number of posts that state that the phase splitting is not as good with this setup, but I have listened to it this way and it sounded good. The other thing would be to revisit the choice of 6SN7 input tube, since the low input level problem goes all the way back to it. I have tried the 5687 in that position and, sorry Chuck, didn't like it as well, but that would be one solution.
Some more thinking to do.
Michael
Hi,
Unfortunately, I didn't get anything done this weekend, so I haven't tested the 6B4G yet. I am concerned about the gain loss as well. Actually, I have been thinking about that since last week when I was listening to some classical piano that was recorded at quite a low level. At max input I could still only get moderate volume out. As I recall from some testing, about 1.1Vrms input drove the 845 to the limit of A1 when it was biased at -125V. I thought that was sufficient gain, as 2Vrms would probably then drive it to the limit of A2, if I could get that working. I was generally trying with this design to have the gain matched to CD player output levels so that minimum attenuation could be used. But that low output level CD kind of upset that thinking. One thing I could do, I think this works, is put a DPDT switch into the secondary side of the LL1676 that would change the input transformer from 1:1+1 to 1:2+2. I have read a number of posts that state that the phase splitting is not as good with this setup, but I have listened to it this way and it sounded good. The other thing would be to revisit the choice of 6SN7 input tube, since the low input level problem goes all the way back to it. I have tried the 5687 in that position and, sorry Chuck, didn't like it as well, but that would be one solution.
Some more thinking to do.
Michael
Hi,
Have you guys condidered a trioded penthode?
A EL34 or something similar should do, I think...
Sorry if it's been suggested before. I haven't kept up with this thread.
Cheers,😉
Have you guys condidered a trioded penthode?
A EL34 or something similar should do, I think...
Sorry if it's been suggested before. I haven't kept up with this thread.
Cheers,😉
Hi Frank,
An EL34 would be a cheap alternative, it also has a quite high gain.
Maybe the internal res is a bit too high to
drive 845 in the positive grid region. I do not know.
Michael what do you think about?
Federico
EL 34 etc
Hi Frank, Federico,
An EL34 could be a possibility, I have a couple I could try. If I read the datasheet right at Ia=70mA, Ra=910 ohms and mu=10.5. I assume Ra goes up as current goes down, so somewhat higher at 50mA max of LL1660. Another possibility I have looked at is the 6BX7. Mu around 10 and Ra=1300 at Ia=42mA. If paralleled sections Ra=650, but somewhat higher at lower Ia/section. I think that Kevin Carter, importer of Lundahl to USA, has built a sort of universal driver circuit around the 6BX7.
To look seperately at the challenges I see at the moment:
1)Gain Structure-may not be a problem (could be fluke low output CD)
a)add switch for hi gain/lo gain on 1676 input transformer.
b)change to input tube with higher mu
c)change to driver with higher mu
2)A2 operation
a)raise plate voltage of 845 and avoid A2 issue
b)change to beefier driver, lower ra, higher Ia
c)add circuit element, i.e. choke loaded cathode follower
Comments or Suggestions on any of these?
Michael
Hi Frank, Federico,
An EL34 could be a possibility, I have a couple I could try. If I read the datasheet right at Ia=70mA, Ra=910 ohms and mu=10.5. I assume Ra goes up as current goes down, so somewhat higher at 50mA max of LL1660. Another possibility I have looked at is the 6BX7. Mu around 10 and Ra=1300 at Ia=42mA. If paralleled sections Ra=650, but somewhat higher at lower Ia/section. I think that Kevin Carter, importer of Lundahl to USA, has built a sort of universal driver circuit around the 6BX7.
To look seperately at the challenges I see at the moment:
1)Gain Structure-may not be a problem (could be fluke low output CD)
a)add switch for hi gain/lo gain on 1676 input transformer.
b)change to input tube with higher mu
c)change to driver with higher mu
2)A2 operation
a)raise plate voltage of 845 and avoid A2 issue
b)change to beefier driver, lower ra, higher Ia
c)add circuit element, i.e. choke loaded cathode follower
Comments or Suggestions on any of these?
Michael
Yes, it works
I have modelled both 2a3 and El34 and they work well.
A part from the different gain, they works similar due to
the high res that loads the drivers (yes 2a3 is a bit more linear but when the load line is so orizzontal the differences are reduced).
They also work well in case of 845 positive grid but this
is a simulation. You remember, also the 12b4a performed well in
simulation but not so in the real circuit in case of grid current.
ciao
Federico
must buckle down
Hi Federico, All,
I have to buckle down and get back to work on this project. I haven't done a damn thing since a week or more. Probably on Friday and this weekend I will get to it again. I will build the actual power supply first so that I can see what kind of voltages I have to play with given choke input and small cap input.
Details, details
Michael
Hi Federico, All,
I have to buckle down and get back to work on this project. I haven't done a damn thing since a week or more. Probably on Friday and this weekend I will get to it again. I will build the actual power supply first so that I can see what kind of voltages I have to play with given choke input and small cap input.
Details, details
Michael
Hi,
The 6BX7 certainly is a good candidate for this kind of service.
Many years ago a good friend if mine developped a 845 PP A2 amp from scratch and as we had the possibility of using any tube in the book we certainly considered about all possible alternatives.
He finally settled for the 6BX7 as it seemed to offer the most potential and he insisted on using nothing but triodes.
Keep in mind that this was to drive the 845s directly as in those days IS xformers were very hard to come by and tubes were much cheaper anyway.
Other alternatives under consideration were the 6BL7s.
2A3s were out as in those days these were too hard to obtain NOS.
We test ran the amps on a pair of Altecs with very good results. In fact the owner of the speakers claimed that all the usual colourations he heard through these speakers had somehow vanished...
Personally, I'd drop the A2 idea as it often presents more of an engineering challenge than the extra Watt or two are worth IMHO.
OTOH, what's life without challenge!?
Cheers and best of luck,😉
The 6BX7 certainly is a good candidate for this kind of service.
Many years ago a good friend if mine developped a 845 PP A2 amp from scratch and as we had the possibility of using any tube in the book we certainly considered about all possible alternatives.
He finally settled for the 6BX7 as it seemed to offer the most potential and he insisted on using nothing but triodes.
Keep in mind that this was to drive the 845s directly as in those days IS xformers were very hard to come by and tubes were much cheaper anyway.
Other alternatives under consideration were the 6BL7s.
2A3s were out as in those days these were too hard to obtain NOS.
We test ran the amps on a pair of Altecs with very good results. In fact the owner of the speakers claimed that all the usual colourations he heard through these speakers had somehow vanished...
Personally, I'd drop the A2 idea as it often presents more of an engineering challenge than the extra Watt or two are worth IMHO.
OTOH, what's life without challenge!?
Cheers and best of luck,😉
A2 engineering
Hi Frank,
I admit I am thinking about giving up the A2 idea, but then I am an engineer, chemical not electrical, so it does rather intrigue me to make a design that works.
I think the 6BX7 has possibilities. The problems I see are that if one tube is used, the plate dissipation is the same as the 12B4 and Ra is higher. Do gain sensitivity. If parallel plates are used Ra is lower than 12B4, but you may drop into a less linear op area. With transformer loading this probably is no problem in A1, but when Rp goes down at A2 ???.
Of course many of the problems I am facing come from self imposed constraints, i.e. desire to build with parts I have on hand, least expense parts, lowest B+, etc. As you said, an interesting design problem.
Michael
Hi Frank,
I admit I am thinking about giving up the A2 idea, but then I am an engineer, chemical not electrical, so it does rather intrigue me to make a design that works.
I think the 6BX7 has possibilities. The problems I see are that if one tube is used, the plate dissipation is the same as the 12B4 and Ra is higher. Do gain sensitivity. If parallel plates are used Ra is lower than 12B4, but you may drop into a less linear op area. With transformer loading this probably is no problem in A1, but when Rp goes down at A2 ???.
Of course many of the problems I am facing come from self imposed constraints, i.e. desire to build with parts I have on hand, least expense parts, lowest B+, etc. As you said, an interesting design problem.
Michael
Maybe yes, grataku, thank you!
Apart from the choke loaded input stage
it appears very similar to Michael topology.
I will try to model 6BX7.
Federico
Apart from the choke loaded input stage
it appears very similar to Michael topology.
I will try to model 6BX7.
Federico
6BX7
Hi grataku, Federico,
I am a little familiar with Kevin's circuit as I ran across it on his website while going through the process of ordering the Lundahl transformers. I did have a brief email discussion with him regarding the alternative merits of the 12B4 and the 6BX7. This was before the project morphed into an 845/A2 project. At that time I was thinking more along the lines of a 2A3 output stage. Looking at Kevin's schematic again I found it interesting that I also found a sweet spot within a few ohms of 220 on the shared cathode resistor. Coincidence?
As I noted in an earlier post the plate resistance of the 6BX7 would seem to be higher than that of the 12B4, which may mean something to successful A2 operation.
On a somewhat unrelated note, last night I found a laboratory kit from the DeForest Training Institute. It was part of their course in Radio, Television and Sound Repair from the late '40s. Included were a few sockets as shown. Great for breadboarding? Don't need clip leads, just pinch the wire or component leads into the appropriate pin connection. Will have to try with EL34 et al this weekend.
Still slogging along,
Michael
Hi grataku, Federico,
I am a little familiar with Kevin's circuit as I ran across it on his website while going through the process of ordering the Lundahl transformers. I did have a brief email discussion with him regarding the alternative merits of the 12B4 and the 6BX7. This was before the project morphed into an 845/A2 project. At that time I was thinking more along the lines of a 2A3 output stage. Looking at Kevin's schematic again I found it interesting that I also found a sweet spot within a few ohms of 220 on the shared cathode resistor. Coincidence?
As I noted in an earlier post the plate resistance of the 6BX7 would seem to be higher than that of the 12B4, which may mean something to successful A2 operation.
On a somewhat unrelated note, last night I found a laboratory kit from the DeForest Training Institute. It was part of their course in Radio, Television and Sound Repair from the late '40s. Included were a few sockets as shown. Great for breadboarding? Don't need clip leads, just pinch the wire or component leads into the appropriate pin connection. Will have to try with EL34 et al this weekend.
Still slogging along,
Michael
Attachments
Glad to see this project is still going!
I've been away for a long long time. Finally landed a job. A nice one too
with a 22 minute commute, potential to bike to work and shower onsite
and pay at about 6x my monthly expense run rate. 🙂
Interesting that you are headed down the road of low mu, low Ra power
triodes... I am going with the 6V6GT triode strapped which is mu=9, Ra=1600.
My bias point will be around 25mA and I'm doing away with IT coupling and
going for LC coupled. Have some really nice plate load chokes. I'm
going to use a 33K grid lead between fixed bias supply and the LC coupling-845
grid node. Any thoughts on that? Im SE BTW- you guys are talking about PP.
-- Jim
I've been away for a long long time. Finally landed a job. A nice one too
with a 22 minute commute, potential to bike to work and shower onsite
and pay at about 6x my monthly expense run rate. 🙂
Interesting that you are headed down the road of low mu, low Ra power
triodes... I am going with the 6V6GT triode strapped which is mu=9, Ra=1600.
My bias point will be around 25mA and I'm doing away with IT coupling and
going for LC coupled. Have some really nice plate load chokes. I'm
going to use a 33K grid lead between fixed bias supply and the LC coupling-845
grid node. Any thoughts on that? Im SE BTW- you guys are talking about PP.
-- Jim
6V6 triode
Jim,
Glad to here about the job. Is it in chemistry? Any other openings there?
I really like the 6V6 triode. The first amp I scratch built was based on it. LC coupling won't work well for A2 if you are considering that mode of operation. Are you thinking on raising the B+ on your 845?
Michael
Jim,
Glad to here about the job. Is it in chemistry? Any other openings there?
I really like the 6V6 triode. The first amp I scratch built was based on it. LC coupling won't work well for A2 if you are considering that mode of operation. Are you thinking on raising the B+ on your 845?
Michael
- Status
- Not open for further replies.