I want to build a single ended KT88 amp and/or an 807. In the UK I can source the Hammond 1628SEA and the Edcore CXSE25-8-5K. I've seen some discussion of these two. There have been some disparaging remarks about the Hammond's top end. If anyone has experience of these I'd appreciate hearing about it. I would have liked to chose the Hammond as they have several output windings/taps, that give matching flexibility and are considerably cheaper than the Edcors. However I don't want to sacrifice sound quality if there is a big difference. Cheers. Matt
The 11 pound Hammond 1628SE 5k single ended can take lots of current. It goes below 10Hz.
But it does have an early high frequency roll off.
I do not know if there are any differences between the 1628SE and 1628SEA
But at 5 or 6 pounds, the Electra Print 5k single ended, and the One Electron 4.8k UBT-2 do not have an early high frequency rolloff.
UBT-2 are hard to get, if they have not already become Un-Obtanium.
Just my experience when I measured all 3 models with the Rohde & Schwarz Vector Network analyzer. It tells the truth, as per the above. I no longer have access to that $50,000 vector network analyzer and $10,000 Calibration kit.
Admittedly, I only listened to the Electra Print and One Electron transformers. They sound great.
I never heard the Hammond 5k, only measured it.
My smaller 10k Hammond push pull, and 10k Edcor push pull work great, in both measurement tests and listening tests.
No high frequency roll off on those.
Now I have to get busy to build; test, and listen to the small to medium size Hammond 5k push pull transformer that I have.
But it does have an early high frequency roll off.
I do not know if there are any differences between the 1628SE and 1628SEA
But at 5 or 6 pounds, the Electra Print 5k single ended, and the One Electron 4.8k UBT-2 do not have an early high frequency rolloff.
UBT-2 are hard to get, if they have not already become Un-Obtanium.
Just my experience when I measured all 3 models with the Rohde & Schwarz Vector Network analyzer. It tells the truth, as per the above. I no longer have access to that $50,000 vector network analyzer and $10,000 Calibration kit.
Admittedly, I only listened to the Electra Print and One Electron transformers. They sound great.
I never heard the Hammond 5k, only measured it.
My smaller 10k Hammond push pull, and 10k Edcor push pull work great, in both measurement tests and listening tests.
No high frequency roll off on those.
Now I have to get busy to build; test, and listen to the small to medium size Hammond 5k push pull transformer that I have.
Thanks for your reply 6A3sUMMER. I did see advice not to use the Hammond SE, but rather use the SEA. The SE having some weirdness at the top end. I'm not sure if this has tainted the reputation of the SEA or whether it is also problems at the top. The spec states +-1dB 20Hz to 20kHz. I may request a frequency/power curve from Hammond.
All this talk about a drop at the high end without any numbers is pretty worthless.
How much does it drop - 0.25dB at 20k? 6dB at 10k? In which application?
Right now we're all groping around in the dark, a hell of a way to make an informed decision.
Jan
How much does it drop - 0.25dB at 20k? 6dB at 10k? In which application?
Right now we're all groping around in the dark, a hell of a way to make an informed decision.
Jan
Last edited:
Ha, well said Jan! The transformer manufactures seem reluctant to state too much about their products. No figures for leakage inductance or power response curves etc. Sowter in UK are quite good, but too expensive for me.
Update! I found a Hammond data sheet that somehow eluded me previously. Apologies to Hammond. Here's the link: https://www.hammfg.com/files/parts/pdf/1628SEA.pdf
Looking at the above data sheet compared to the original one I looked at: https://www.mouser.co.uk/datasheet/2/177/1627_1642-1389902.pdf
I see a couple of anomalies. One shows the frequency response +-1dB the other +-1.5dB. Not significant. However, the primary inductance is stated as 48Henrys in one spec and 16.2 Henrys @1kHz, 1.0V, Open circuit, in the other. I imagine the 48H was measured under different conditions as this is very different. I can't find similar information on the Edcor, but have requested it from them.
I see a couple of anomalies. One shows the frequency response +-1dB the other +-1.5dB. Not significant. However, the primary inductance is stated as 48Henrys in one spec and 16.2 Henrys @1kHz, 1.0V, Open circuit, in the other. I imagine the 48H was measured under different conditions as this is very different. I can't find similar information on the Edcor, but have requested it from them.
The SE was a train wreck. Hammond's unique secondary winding config in 8 ohms resulted in a resonant notch around 13 kHz. This was reported corrected in the SEA version.
The original 1628SE had some well-documented issues:
https://www.diyaudio.com/community/threads/what-s-wrong-with-hammond-1628se.29873/
http://tubelab.com/articles/component-testing/output-transformers/
I'm pretty sure I posted a frequency response graph back in the day. The impedance graph below of a first gen 1628SE was recently generated with a DATS driving the 5K primary with an 8 ohm load on the output using Hammond's recommended winding connections. No DC primary current.
Thanks. That curve is very different from the Hammond 1628SEA, which rolls off smoothly at the top. Tubelab says the SEA is curtailed at the top compared to the Edcor and that the size of its core means a magnetic inertia that can destroy low level information......
I will see if I can find the Rhode & Schwarz plot I made of the 1625SE.
In the meantime, there are Marketing tricks.
Which is better for an amplifier that uses an output transformer?
1. + / - 1 dB from 20Hz to 20kHz
Or,
2. +0 / - 2dB from 20Hz to 20kHz
. . . well the above two frequency responses are typically the same, because most amplifiers do not have peaks in the middle of the audio band.
If they do have a peak in the middle of the audio band, they need to be re-designed or fixed.
So, for this example, the amplifier will be 0dB at 1kHz, and it will be -2dB at 20Hz, and -2dB at 20kHz.
It rolls off at both frequency extremes, but does not have any peak in the middle.
And that same response curve can also can be specified to be + / - 1 dB from 20Hz to 20kHz.
Do you see the trick?
Your mind overlooks a +1 dB variation and a -1dB variation, but in reality the total variation is 2 dB, and your mind does not like 2dB ("too large for you").
As a former marketeer who competed with other creative marketeers, I know a trick or two.
In the meantime, there are Marketing tricks.
Which is better for an amplifier that uses an output transformer?
1. + / - 1 dB from 20Hz to 20kHz
Or,
2. +0 / - 2dB from 20Hz to 20kHz
. . . well the above two frequency responses are typically the same, because most amplifiers do not have peaks in the middle of the audio band.
If they do have a peak in the middle of the audio band, they need to be re-designed or fixed.
So, for this example, the amplifier will be 0dB at 1kHz, and it will be -2dB at 20Hz, and -2dB at 20kHz.
It rolls off at both frequency extremes, but does not have any peak in the middle.
And that same response curve can also can be specified to be + / - 1 dB from 20Hz to 20kHz.
Do you see the trick?
Your mind overlooks a +1 dB variation and a -1dB variation, but in reality the total variation is 2 dB, and your mind does not like 2dB ("too large for you").
As a former marketeer who competed with other creative marketeers, I know a trick or two.
Years ago SAC-Thailand tested various OPTs,
The Hammond 1628SE was not good:
http://www.sacthailand.com/transformer_testoutput.html
http://www.sacthailand.com/Transformer_OutputTestGraph.html
The Hammond 1628SE was not good:
http://www.sacthailand.com/transformer_testoutput.html
http://www.sacthailand.com/Transformer_OutputTestGraph.html
Useful info FullRangeMan. Thanks. Hammond seem to have addressed the top end nasty issue in the SEA. They have also reduced the copper losses from 439R t0 171R. The leakage inductances look a bit high though.
My misgivings about the SEA have been sown by Tubelab. Unfortunately the Edcor has only one output tap and would cost me over £100 more than the Hammond.
My misgivings about the SEA have been sown by Tubelab. Unfortunately the Edcor has only one output tap and would cost me over £100 more than the Hammond.
If you are in the UK, I have a pair of unused Hammond 1640SEA 30watt OPTs for sale at just over half the new price + postage.
I had planned to build a parallel SE amp with four KT120s but have decided to stick with soild state (safer when you have shaky hands!)
That sounds good. I'm in south east London. Can you give me a total price including shipping? Many thanks batteryman.
Oh, sorry, I got over excited, and misread 1640 as 1628, so not quite what I'm after. Thanks anyway.