Hi all 😀
I recently completed a parafeed EL84 headphone amplifier. I chose it because I had all the parts and was looking for a simple, fun tube project to build over the holidays.
The tube circuit is a clone of Jim Hagerman's Tuba. I don't want to republish the full schematic here, but it can be found in the user manual: http://www.hagtech.com/pdf/tuba.pdf
I built it point to point and it powers up fine, absolutely no hum. All the voltages check out. However, there is nasty low frequency roll off. The output transformer is a speaker matching style unit - 600ohm primary with 8ohm secondary and a 4ohm centre tap.
Where should I start looking to address the frequency response issue? Is it possibly poor transformers?
Signal schematic and FR plot are attached.
I recently completed a parafeed EL84 headphone amplifier. I chose it because I had all the parts and was looking for a simple, fun tube project to build over the holidays.
The tube circuit is a clone of Jim Hagerman's Tuba. I don't want to republish the full schematic here, but it can be found in the user manual: http://www.hagtech.com/pdf/tuba.pdf
I built it point to point and it powers up fine, absolutely no hum. All the voltages check out. However, there is nasty low frequency roll off. The output transformer is a speaker matching style unit - 600ohm primary with 8ohm secondary and a 4ohm centre tap.
Where should I start looking to address the frequency response issue? Is it possibly poor transformers?
Signal schematic and FR plot are attached.
Attachments
Last edited:
First try to increase the coupling capacitor 4,7uF ... for testing you can use electrolytics . 600ohm is a low load , so the coupling cap should be pretty big
Last edited:
Are you use Hammond 157G choke and 119DA OPT as in the original Hagerman?
https://nz.element14.com/hammond/119da/transformer-600ohm-8ohm-12w/dp/2302188
https://nz.element14.com/hammond/119da/transformer-600ohm-8ohm-12w/dp/2302188
Attachments
Last edited:
You will have a -3db roll off frequency of ~56hz with a ~600ohm load and a 4.7uF capacitor
That makes sense, but doesn't explain how early the roll-off starts. I'll try a bigger cap and see how much it helps 🙂.
Too expensive here unfortunately, but I had a well-regarded local transformer winder make me equivalents.
"make me equivalents"
With these parameters?
https://www.google.hu/url?sa=t&rct=...df/119DA.pdf&usg=AOvVaw2Ua-Cj2CjFIAR1VY7V5mpa
With these parameters?
https://www.google.hu/url?sa=t&rct=...df/119DA.pdf&usg=AOvVaw2Ua-Cj2CjFIAR1VY7V5mpa
Normaly when using 32ohm headphones hooked at the 8ohm tap the load is 2K4 , or 4K8 with the 4ohm tap . We don't know how you test it . 600ohm is to low for normal operation of the tube .
I don't think this circuit provides a low 600 Ohm output, must be a cathode follower to do so for driving loudspeakers.
So I think its a mismatch between output impedance of circuit and transformer input impedance of parafeed output.
So I think its a mismatch between output impedance of circuit and transformer input impedance of parafeed output.
Yup, I sent them the Hammond data sheets.
The plot I posted is for a 32ohm load on the 4ohm tap.
I initially thought the same, but it comes down to the effective load the tube sees I believe? For example, the Papa Rusa article (Papa Rusa Headphone Amplifier – wauwatosa tube factory) says the same Hammond OPT would work as an alternative to an Edcor 10k:150. I am open to correction here though, as tubes are not my forte.
None of the reviews I read mentioned poor low frequency response - when listening it is not subtle at all. Which makes me think my particular OPT is the culprit...
Last edited:
I swapped out the 4.7uF coupling cap in one channel for a 47uF one. The plot actually looks a little worse - an additional 3 dBV down at 100Hz, a hump at 50Hz and a very small hump at 106Hz. Hmmm.
Your output is generally not a low impedance one. But every audio signal transformer is just a mirror for the impedances, which were reflected with the winding proportions that transformer has been designed and build to. So the Wauwatosa- Edcor can reflect any 150 Ohm load to a 10K load that the tube sees. If its being used with a lower impedance headphone, a mismatch will occur and the trannie will reflect a smaller impedance to the circuit.
But still this wrong impedance is way bigger than a 600 Ohm primary transformer, I would say it becomes a 2K load on primary at one fifth the secondary load impedance.
But still this wrong impedance is way bigger than a 600 Ohm primary transformer, I would say it becomes a 2K load on primary at one fifth the secondary load impedance.
Last edited:
This takes quite a bit of a long story, in order to get to the answer.
A triode wired EL84 has a plate resistance, rp, of about 1700 Ohms, the same as a type 45 triode.
A 600 Ohm output transformer, that is 600:8 Ohms works fine, when the rp is about 200 Ohms, and the load is 8 Ohms (or an rp of no more than 600 Ohms, which will distort at bass frequencies).
Instead, you are driving with an rp of about 1700 Ohms, and the load is 32 Ohms.
That means the transformer is now 32/8 x 600 Ohms = 2400 Ohm primary.
That is 'almost' OK.
The output transformer for a 45 rp of 1700 Ohms is 4600 Ohms ( that is lots more than the 2400 Ohms above).
But the real detail that gets you is this . . .
A 4600 Ohm transformer has enough inductance so that the inductive reactance is equal to 4600 Ohms at the -3dB point of the transformers low frequency rated bandwidth.
At 20 Hz, the inductive reactance of 4600 Ohms, requires 36.6 Henrys.
At 40 Hz, the inductive reactance of 4600 Ohms, requires 18.3 Henrys.
Does your output transformer have at least 18, or 36 Henrys?
If not, there goes the low frequency response.
Oh, I forgot, your choke is only 30 Henrys. That 30 Henrys is in parallel with your output transformer inductance.
If your output transformer is 30 Henrys, the total parallel inductance is 15 Henrys.
15 Henrys and 2400 Ohms is - 3dB at 25.5 Hz. Just remember, you are driving 2400 Ohms with an rp of 1700 Ohms, that is not ideal.
And, 4.7uF is 1693 Ohms at 20 Hz, another cause for low frequency roll off.
A triode wired EL84 has a plate resistance, rp, of about 1700 Ohms, the same as a type 45 triode.
A 600 Ohm output transformer, that is 600:8 Ohms works fine, when the rp is about 200 Ohms, and the load is 8 Ohms (or an rp of no more than 600 Ohms, which will distort at bass frequencies).
Instead, you are driving with an rp of about 1700 Ohms, and the load is 32 Ohms.
That means the transformer is now 32/8 x 600 Ohms = 2400 Ohm primary.
That is 'almost' OK.
The output transformer for a 45 rp of 1700 Ohms is 4600 Ohms ( that is lots more than the 2400 Ohms above).
But the real detail that gets you is this . . .
A 4600 Ohm transformer has enough inductance so that the inductive reactance is equal to 4600 Ohms at the -3dB point of the transformers low frequency rated bandwidth.
At 20 Hz, the inductive reactance of 4600 Ohms, requires 36.6 Henrys.
At 40 Hz, the inductive reactance of 4600 Ohms, requires 18.3 Henrys.
Does your output transformer have at least 18, or 36 Henrys?
If not, there goes the low frequency response.
Oh, I forgot, your choke is only 30 Henrys. That 30 Henrys is in parallel with your output transformer inductance.
If your output transformer is 30 Henrys, the total parallel inductance is 15 Henrys.
15 Henrys and 2400 Ohms is - 3dB at 25.5 Hz. Just remember, you are driving 2400 Ohms with an rp of 1700 Ohms, that is not ideal.
And, 4.7uF is 1693 Ohms at 20 Hz, another cause for low frequency roll off.
Last edited:
Problem is, that this output circuit was never designed to drive 600 Ohm loads.
If you want to do so, use one of the many cathode follower circuits (or LA2A output section, WCF etc.). Of course, there are those low imp. tubes that could do so, but in practice, with ordinary tubes, I don't think this will give good results.
Western Electric did it with a 16K interstage, and that seem to work properly, because its not such a heavy load on the tube.
If you want to do so, use one of the many cathode follower circuits (or LA2A output section, WCF etc.). Of course, there are those low imp. tubes that could do so, but in practice, with ordinary tubes, I don't think this will give good results.
Western Electric did it with a 16K interstage, and that seem to work properly, because its not such a heavy load on the tube.
A little better with lower Ra tubes but the primary inductance is too low ( only 3,5H ) . This should be enough for true 600ohm load but not for tubes , at least not this kind of small tubes and in this configuration .
But there are plenty of cheap SE transformers for EL84 that after removing the air gap should have good inductance .
If the transformer used has also air gap than it could be made to perform better ...
But there are plenty of cheap SE transformers for EL84 that after removing the air gap should have good inductance .
If the transformer used has also air gap than it could be made to perform better ...
Last edited:
That's extremely useful, thank you! So if I have a 32 ohm load on the 4 ohm tap the primary is then 4800 ohms. The math tells me this would need 19H for 40Hz -3dB, but with the 30H in parallel that becomes a whopping 52H needed. Which I doubt I have. But I also very much doubt the Hammond unit Hagerman uses is that high either - it's too small, physically. I'm going to measure my transformer now.
Tubes are hard to get here, and when one can they are very expensive. I had EL84s on hand. I Googled "EL84 headphone amplifier" and one of the circuits I found was Hagerman's
I wanted something simple that could be made without too much expense. I came across the Hagerman design, and since he's well respected I thought I'd give it a go.
I know the 30H inductor I have is gapped, will check the OPT.
Thank you for the input, everyone! You guys rock 😀