Or the other way round, I have the new resistors for the 2A3, I won't get the parts for the ECC99 driver until some time next week. I'm also waiting on parts for the current regulated 2A3 filaments, I've got partially soldered boards sitting waiting.
But you're right, I'm fairly familiar with the sound of the amps as they are now, I should wait.
That's true.
I do have Sovteks, and I'm at about 250V/55mA right now. Frank tried it at a higher dissipation and said he liked it. No one really knows what that does to the longevity though.
But you're right, I'm fairly familiar with the sound of the amps as they are now, I should wait.
That's true.
I do have Sovteks, and I'm at about 250V/55mA right now. Frank tried it at a higher dissipation and said he liked it. No one really knows what that does to the longevity though.
Hey-Hey!!!!,
I just had to jump in on this one. I am a big fan of the cascoded MOSFET CCS. Preferably battery biased and built around the DN2540/DN3545 devices. These are depletion mode and for a plate load offer a few advantages over the C4S and similar.
Connect the load to the lower source of the CCS, this way the tube sees a near-infinite load and the output z is that of a high gm mu follower, or put basically, in the hundreds of ohms, instead of the tube's plate resistance as what would be seen with the 'standard' use of a CCS, connected to the other side of the curent set resistor.
For your project a 10 mA setting would be more than adequate, given the so called 'mu' connection as taken form Gary Pimm's site. The lower device will do what it can to keep the voltage across the set resistor constant, and any curent required by the load will be supplied without upsetting the prime benefit of the CCS, that is a horizontal load line.
Look at he ECC99 curves, they do not change much more after 10 mA. that is the slope of the individual grid voltage lines is nearly constant after the 10 mA line has been crossed. This would be of more concern if you use a C4S or connect the load to the plate, but with the 'mu' connection, you get a free lunch almost...BTW, increase the size of the coupling cap, to something like 1 uF of good cap.
regards,
Douglas
I just had to jump in on this one. I am a big fan of the cascoded MOSFET CCS. Preferably battery biased and built around the DN2540/DN3545 devices. These are depletion mode and for a plate load offer a few advantages over the C4S and similar.
Connect the load to the lower source of the CCS, this way the tube sees a near-infinite load and the output z is that of a high gm mu follower, or put basically, in the hundreds of ohms, instead of the tube's plate resistance as what would be seen with the 'standard' use of a CCS, connected to the other side of the curent set resistor.
For your project a 10 mA setting would be more than adequate, given the so called 'mu' connection as taken form Gary Pimm's site. The lower device will do what it can to keep the voltage across the set resistor constant, and any curent required by the load will be supplied without upsetting the prime benefit of the CCS, that is a horizontal load line.
Look at he ECC99 curves, they do not change much more after 10 mA. that is the slope of the individual grid voltage lines is nearly constant after the 10 mA line has been crossed. This would be of more concern if you use a C4S or connect the load to the plate, but with the 'mu' connection, you get a free lunch almost...BTW, increase the size of the coupling cap, to something like 1 uF of good cap.
regards,
Douglas
Bandersnatch said:
Me too. These are what I'm going to use for my CCS referred to in the earlier posts. Because I don't have some of the parts required on hand, they'll be singles (not cascoded) for a while.
This is what I thought too until I tried it. Not a huge difference, but significant enough, and farly constant across all the '99's I have.
Cascode is far better than the single
I could barely see a diff in the loadline between a Meg and several or several hundred, but the sound, the sound is far more lifelike and pleasing with the low capacitance cascode circuit. Just knnow that the 2540 wants something like 500 ohm gate stopper/dampers.
have fun with them, and try it all for yourself, you can convince you far easier than I can
regards,
Douglas
I could barely see a diff in the loadline between a Meg and several or several hundred, but the sound, the sound is far more lifelike and pleasing with the low capacitance cascode circuit. Just knnow that the 2540 wants something like 500 ohm gate stopper/dampers.
have fun with them, and try it all for yourself, you can convince you far easier than I can
regards,
Douglas
I've seen that done on a couple of other circuits. I'll keep that in mind. For now, I wanted a kit where someone else had done the figuring out part for me.
Hi Saurav,
K&K audio sells a CCS kit with the DN2540 mosfet..(If you can call it a kit ) because of simplicity...🙂
1 resistor, 1 DM mosfet, 1 circuitboard, 1 pot. Setting the right current is even easier than with a C4S.
The kit comes down to the simple CCS I have on my CCS page.
Only difference is the gate stopper is a little higher in my simple CCS. The more complex CCS's are Gary Pimm/Douglas inspired and are up on my site for reference.
http://home.zonnet.nl/horneman/mosfet.htm
Cheers,
Bas
K&K audio sells a CCS kit with the DN2540 mosfet..(If you can call it a kit ) because of simplicity...🙂
1 resistor, 1 DM mosfet, 1 circuitboard, 1 pot. Setting the right current is even easier than with a C4S.
The kit comes down to the simple CCS I have on my CCS page.
Only difference is the gate stopper is a little higher in my simple CCS. The more complex CCS's are Gary Pimm/Douglas inspired and are up on my site for reference.
http://home.zonnet.nl/horneman/mosfet.htm
Cheers,
Bas
Hi Douglas, you don't have to convince me about the cascoded CCS/mu. I've been using Gary's BBMCCS for a while now as a mu and the 2540's will be next. The circuit I came up with to try is very similar to the Mk4 version of yours Bas has on his site. Thanks for the tip on the size of the stoppers.
I think you and I are about the only CCS adherents here.
Living in the country means I have to mail order all my parts, so I try to get up a decent size order to make the postage more economical, hence the singles for the time being.
I think you and I are about the only CCS adherents here.
Living in the country means I have to mail order all my parts, so I try to get up a decent size order to make the postage more economical, hence the singles for the time being.
...any old iron...
I have to confess to trying the ccs that Bas has on his site and to trying the mu connection as well. This was in input and driver stages (6SN7) for my sons WAD 2A3 PSE amp. (It's quite nice as standard but could be better. ) Since the first two stages are resistor loaded - moving to ccs loaded was easy. It did improve the sound but not by a lot. I thought it should make a big difference so I tried a couple of plate chokes - now that was much better! I'm just about to get radical with it as my ECL82Se amp is still better - so now I'm going to two stages only and replacing the input and driver with a single Bi valve (DHT mu=27) IT connected to the 2A3s. I'll use the LL1660 as interstage and use one each of the secondaries to drive each parallel 2A3... I bet it the best yet I'll let you know...
But the point is that the ccs did not show the improvement in sound I expected and I'm not sure why??? Any suggestions?
I will try them in other amps 'cos it might just be that one instance that is like this...
ciao
James
I have to confess to trying the ccs that Bas has on his site and to trying the mu connection as well. This was in input and driver stages (6SN7) for my sons WAD 2A3 PSE amp. (It's quite nice as standard but could be better. ) Since the first two stages are resistor loaded - moving to ccs loaded was easy. It did improve the sound but not by a lot. I thought it should make a big difference so I tried a couple of plate chokes - now that was much better! I'm just about to get radical with it as my ECL82Se amp is still better - so now I'm going to two stages only and replacing the input and driver with a single Bi valve (DHT mu=27) IT connected to the 2A3s. I'll use the LL1660 as interstage and use one each of the secondaries to drive each parallel 2A3... I bet it the best yet
I'll let you know...But the point is that the ccs did not show the improvement in sound I expected and I'm not sure why??? Any suggestions?
I will try them in other amps 'cos it might just be that one instance that is like this...
ciao
James
When you replaced the anode load resistor with a CCS, the following grid leak resistor was still in parallel at AC, so you didn't flatten the loadline by as much as you expected. I always feel that if you use a CCS load, you have to follow it with a DC coupled cathode follower to avoid this loading problem. That's one reason why the mu-follower is so nice.
The big question is whether you kept the anode voltage the same when you inserted the choke. Typically, using an inductive load doubles the maximum swing, or for the same swing, halves the distortion. Perhaps that was what happened?
The big question is whether you kept the anode voltage the same when you inserted the choke. Typically, using an inductive load doubles the maximum swing, or for the same swing, halves the distortion. Perhaps that was what happened?
Hi James,
The more I build things the more I realize that certain things are too hyped. For instance...my SE 6c41c sounds better than at least one 300B amp I could directly compare it against in the same setup. "On paper" a 300b should be one of the very best sounding tubes out there.
Same with the CCS..on paper it should kick resistor butt! In real life I suppose it sounds "different" rather than better...some preferring a simple resistor to a CCS...in any case...for many "newbies" like myself...when first trying a CCS...we are shocked (lesson learnt) that it does not sound 100% better!
This is the reason I think that more seasoned builders stop looking for the holy grail / tube of the month / boutique part and just enjoy the building, try to make it look better (prettier) and most importantly try to voice their entire system to be good at what they like.
Cheers,
Bas
Hi EC,
You are absolutely right!!! Infact the driver stage looks into a 150K grid leak and has a 56k anode resistor so the change is worthwhile but not that dramatic. I set the ccs current for just a little more than the resistor loaded stage so the OP didn't move much in this case. The mu connection was better, by about the same as the anode out ccs compared to the resistor loaded stage.
Spot on the money again... I run the choke loaded driver stage at a much hotter OP - over twice the current and another 100V on the anode... Not surprising it sounds different🙂 One of the big advantages of iron in the signal path is the ability to run the valves at hotter OPs. I know it's not apples with apples and I really should go back and try the CCS at the same current...but I haven't had time yet and I want to try the interstage next...
Hi Bas
'ain't that the truth!!!
I have yet to hear a really good (to my ears) 300B... I've tried them in a quick breadboard design of my own too (borrowed valves) and found then too thick and not transparent at all. Nice tone but too much tone if you know what I mean
Yep! End to End design is everything - and that includes the listening environment and the listeners...
ciao
James
You are absolutely right!!! Infact the driver stage looks into a 150K grid leak and has a 56k anode resistor so the change is worthwhile but not that dramatic. I set the ccs current for just a little more than the resistor loaded stage so the OP didn't move much in this case. The mu connection was better, by about the same as the anode out ccs compared to the resistor loaded stage.
Spot on the money again... I run the choke loaded driver stage at a much hotter OP - over twice the current and another 100V on the anode... Not surprising it sounds different🙂 One of the big advantages of iron in the signal path is the ability to run the valves at hotter OPs. I know it's not apples with apples and I really should go back and try the CCS at the same current...but I haven't had time yet and I want to try the interstage next...
Hi Bas
'ain't that the truth!!!
I have yet to hear a really good (to my ears) 300B... I've tried them in a quick breadboard design of my own too (borrowed valves) and found then too thick and not transparent at all. Nice tone but too much tone if you know what I mean
Yep! End to End design is everything - and that includes the listening environment and the listeners...
ciao
James
Hi James,
I am in my lab right now, trying to create two 19 year old females that are as much into tubes as I am and help me listen to my system.
Females have better ears, is the reason ofcourse!
Cheers,
Bas
Frankenstein resurrected
Just remember to leave out the bolt through their necks, then post the design. I doubt if anyone will complain about it being off-topic.
Bas Horneman said:I am in my lab right now, trying to create two 19 year old females that are as much into tubes as I am and help me listen to my system. Bas
Just remember to leave out the bolt through their necks, then post the design. I doubt if anyone will complain about it being off-topic.
If you have two females in the room with you - how do you stop them talking to each other when you are trying to listen???
James D. said:
Thank goodness, I thought it was just me.
Two for two....
But all that tone has it's uses in other places. But shush or we'll have all the lemmings running headlong into another audio fad.
Personally, I want to get another 50 300B's and arrange them in a pattern like a Shakti Hallograph Soundfield Optimiser to impress all my friends.
An externally hosted image should be here but it was not working when we last tested it.
Woo hoo!
RE:300B
Hi,
Driving a 300B with a penthode and heating them with DC helps.
To my ears and some others the 2A3 is much better sounding.
Cheers,😉
P.S. Bas, don't electrocute those 19 year olds, will you?😎
Hi,
Driving a 300B with a penthode and heating them with DC helps.
To my ears and some others the 2A3 is much better sounding.
Cheers,😉
P.S. Bas, don't electrocute those 19 year olds, will you?😎
That too is against forum rules...
That is against the law Frank, at least here in Holland it is!
Hi,
Bas,
My only experience with a CCS was when I replaced an anode resistor with Garry' s BBMCCCS on a 6C45P, in common cathode configuration, which is used as an output amplifier buffer of a TDA1541 DAC. The difference I noticed was quite impressive.
Since I have not used a CCS in another circuit to compare, I can understand that a CCS may not be a panacea and that certain topologies and/or tubes may benefit from a CCS, while with other combinations things may go worse. But I can wholeheartedly say that the BBMCCS proved to be a lot beneficial to the circuit it was used in and that it surely sounded 50% (if not more) better.
Just my 2 pennies.
Evangelos
Bas,
My only experience with a CCS was when I replaced an anode resistor with Garry' s BBMCCCS on a 6C45P, in common cathode configuration, which is used as an output amplifier buffer of a TDA1541 DAC. The difference I noticed was quite impressive.
Since I have not used a CCS in another circuit to compare, I can understand that a CCS may not be a panacea and that certain topologies and/or tubes may benefit from a CCS, while with other combinations things may go worse. But I can wholeheartedly say that the BBMCCS proved to be a lot beneficial to the circuit it was used in and that it surely sounded 50% (if not more) better.
Just my 2 pennies.
Evangelos
- Status
- Not open for further replies.