Haven't seen it, but it don't meet my requirements as it's a hybrid tube/mosfet design.... I want only tubes in the signal path, although I'm ready to use semiconductors elsewhere....
The ECC88 is not powerful enough (others might have another opinion), but do have enough gain for a one stage design.... I want more power so a driver tube is needed, which I haven't been able to find with enough gain....
I like the advantages of the parafeed principle, at cost of just a larger power supply.
I did that when trying first with a single stage design....
I don't want anything to adjust, either I get the current source matched perfectly, or I will add a capacitor between the two primary windings.
Eeeh, 12HL7 and the others are single pentodes and don't seems to be made anymore....
I want to use dual triodes that are easily available, which to me means they are still being manufactured....
But you can probably buy a truckload of 4P1L from Russia for little money... (still).
Probably, but when running a business you can't do with maybe getting a truckload at some time you then have to test and sort....
I like the principle of ordering parts that get delivered at the agreed date, with all new and perfectly working parts.
I would be very interested to hear opinion about tubes and their alternatives, but please note that any alternative to the ECC99 output tube:
Have to be in production
Have to be a dual triode
Have to have anode dissipation of min 4W per triode
Have as low internal plate resistance as possible
So far I have ECC99, 6N6P, 6H30P, 6SN7, 12BH7, 6GC7. Some of those will require higher anode supply that shown in schematics....
Pinouts are not that important, I will have the tubes on a separate PCB, mounted horizontally on the main PCB for a low profile case....
Have to be in production
Have to be a dual triode
Have to have anode dissipation of min 4W per triode
Have as low internal plate resistance as possible
So far I have ECC99, 6N6P, 6H30P, 6SN7, 12BH7, 6GC7. Some of those will require higher anode supply that shown in schematics....
Pinouts are not that important, I will have the tubes on a separate PCB, mounted horizontally on the main PCB for a low profile case....
The E55L would make an excellent single stage single tube candidate and works very well in parafeed. Its powerful enough to drive sensitive speakers to loud levels even ion triode mode. Extremely linear and neutral sounding valve.
Just an observation, if you are going parafeed then SE is the order of the day since it is relatively simple to make a zero DC offset Push Pull amp without any sxpecial measures like parafeed. If I was looking for a simple PP design for headphones I would consider the 5687 which is like a beefy ECC88. Place a phase splitting choke up front and a 2+2:1 transformer at the output and you have just about the most elegant design imaginable. If the 5687 proves inadequate the same setup with those E55L would drive any headphone to melting point.
Shoog
Just an observation, if you are going parafeed then SE is the order of the day since it is relatively simple to make a zero DC offset Push Pull amp without any sxpecial measures like parafeed. If I was looking for a simple PP design for headphones I would consider the 5687 which is like a beefy ECC88. Place a phase splitting choke up front and a 2+2:1 transformer at the output and you have just about the most elegant design imaginable. If the 5687 proves inadequate the same setup with those E55L would drive any headphone to melting point.
Shoog
Hi Soekris,
With current production I think you are stuck with ECC99. Maybe 6N30P (i think they are being manufactured...). 12BH7A is a nice tube, but will need an additional gain stage, with ECC99 you may be able to skip that.
Just some questions:
- what output power do you want? I ask because there is some headphones that like 5W, I am sure you don't want 5W of your amp, but a figure would be nice.
- what is the required input sensitivity for full (desired) power? The raw 1,6VRMS (SE) of your DAM1021? Or you are planning an opamp in between for balanced outputs that could be used for the headphone amp as well?
I am a big fan of the OPT arrangement with multiple secondaries that can be connected in series/parallel for multiple impedances: specially in the headphone world this is very nice to have. For example 4x 16R output windings could be wired for 16R, 64R, ~150R and 256R. I am sure you will find some electronic switch that could do this switching.
HEADPHONE TRANSFORMERS
For parafeed I think a good option may be the use of a gyrator per side of the PP output stage: the gyrators will keep a constant voltage at both outputs and maybe you can skip the DC blocking cap for the OPT. There is a lot of recent activity with gyrators here at diyaudio, look at posts from Mogliaa, MrCurwen, etc.
if you have access to Morgan Jones 4th edition, he writes about the "harmonizer resistor", a resistor shared by both cathodes of the output tubes that influences the harmonic distribution of the output stage. Would be nice to include something like that in this amp, maybe even adjustable so clients can adjust for best distortion with their specific headphone.
good luck with the project!
Erik
With current production I think you are stuck with ECC99. Maybe 6N30P (i think they are being manufactured...). 12BH7A is a nice tube, but will need an additional gain stage, with ECC99 you may be able to skip that.
Just some questions:
- what output power do you want? I ask because there is some headphones that like 5W, I am sure you don't want 5W of your amp, but a figure would be nice.
- what is the required input sensitivity for full (desired) power? The raw 1,6VRMS (SE) of your DAM1021? Or you are planning an opamp in between for balanced outputs that could be used for the headphone amp as well?
I am a big fan of the OPT arrangement with multiple secondaries that can be connected in series/parallel for multiple impedances: specially in the headphone world this is very nice to have. For example 4x 16R output windings could be wired for 16R, 64R, ~150R and 256R. I am sure you will find some electronic switch that could do this switching.
HEADPHONE TRANSFORMERS
For parafeed I think a good option may be the use of a gyrator per side of the PP output stage: the gyrators will keep a constant voltage at both outputs and maybe you can skip the DC blocking cap for the OPT. There is a lot of recent activity with gyrators here at diyaudio, look at posts from Mogliaa, MrCurwen, etc.
if you have access to Morgan Jones 4th edition, he writes about the "harmonizer resistor", a resistor shared by both cathodes of the output tubes that influences the harmonic distribution of the output stage. Would be nice to include something like that in this amp, maybe even adjustable so clients can adjust for best distortion with their specific headphone.
good luck with the project!
Erik
Hi Erik,
The 6N30P is still being manufactured, but the ECC99 have more making them and are lower cost.... I manage to get 6N30P working in ltspice, it's lower voltage / impedance than ECC99, but otherwise much alike.
The 12BH7A is higher voltage / impedance than ECC99, but also otherwise much alike.
As someone else pointed out, the 5687 is like a little beefier ECC99 and that would be perfect, but I can't find any new manufacturers of the 5687....
When I'm comparing tubes it is current from an objective view, the subjective sound is a completely different topic....
I'm targeting 0.5W, but it will be more linear at 0.25W. 0.5W is 11V into 250 ohm, for a typical highend sennheiser, should be plenty....
Input sensitivity down to 1V, need to be able to be driven by an unbuffered dam1021.... So a two stage design is needed. Will probably have a switch for high / low sensitivity. Volume potmeter will be optional.
I have been simulation different loads having two or four output secondaries, and have so far decided that two is enough, connected parallel or serial by switch.
Will look into that, but I'm worried about AC performance of such circuit, don't want any NFB.... But would be nice to avoid capacitors.
Don't have any tube books at all, will look into it and probably visit amazon....
Thanks, and thanks for the good feedback.
The 6N30P is still being manufactured, but the ECC99 have more making them and are lower cost.... I manage to get 6N30P working in ltspice, it's lower voltage / impedance than ECC99, but otherwise much alike.
The 12BH7A is higher voltage / impedance than ECC99, but also otherwise much alike.
As someone else pointed out, the 5687 is like a little beefier ECC99 and that would be perfect, but I can't find any new manufacturers of the 5687....
When I'm comparing tubes it is current from an objective view, the subjective sound is a completely different topic....
I'm targeting 0.5W, but it will be more linear at 0.25W. 0.5W is 11V into 250 ohm, for a typical highend sennheiser, should be plenty....
Input sensitivity down to 1V, need to be able to be driven by an unbuffered dam1021.... So a two stage design is needed. Will probably have a switch for high / low sensitivity. Volume potmeter will be optional.
I have been simulation different loads having two or four output secondaries, and have so far decided that two is enough, connected parallel or serial by switch.
Will look into that, but I'm worried about AC performance of such circuit, don't want any NFB.... But would be nice to avoid capacitors.
Don't have any tube books at all, will look into it and probably visit amazon....
Thanks, and thanks for the good feedback.
He spoke well. Any triode will give significantly lower thd, when sees high load impedance. Gain is very close to µ.
My 2A3´s see up to 17k(!) and absolute none FB, sounds great. "standard" load is 2k5 and FB..
The reality is that this low THD cannot be realized in a practical circuit since the headphone load totally swamps it. We should be using CCS and gyrators in this application for their other benefits (ie DC balance).He spoke well. Any triode will give significantly lower thd, when sees high load impedance. Gain is very close to µ.
My 2A3´s see up to 17k(!) and absolute none FB, sounds great. "standard" load is 2k5 and FB..
Shoog
The secondary should be able to deliver to loads between 32 and 600R. That is a lot of variation. The critical thing is to remember the 'amp' portion of this 'headphone amp' requirement. It is just an amp, and if you can exclude Grado's with their 32R load things get easier.
I dislike the parafeed idea; I do like the SE exclusion...LOL Do an OPT core with EI, and build in a *SMALL* gap and you can leave current imbalance out of the picture. I am sure a small stack of low NI lams with maybe a .05mm gap will do quite well.
cheers,
Douglas
I dislike the parafeed idea; I do like the SE exclusion...LOL Do an OPT core with EI, and build in a *SMALL* gap and you can leave current imbalance out of the picture. I am sure a small stack of low NI lams with maybe a .05mm gap will do quite well.
cheers,
Douglas
Yes, you're right. That LM317 should have the ADJ grounded and the resistor from output to ground. I just grabbed this schematic from an old post and hadn't ever fixed the image.
I put together the schematic for someone that had not built anything with tubes before and used the simplest CCS I could think of. Certainly could substitute a couple of BJTs or other CCS if so desired.
I put together the schematic for someone that had not built anything with tubes before and used the simplest CCS I could think of. Certainly could substitute a couple of BJTs or other CCS if so desired.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.