Leakage inductance tests
I did some tests on the leakage inductance of the 1628SE and 1640SE versus the secondary winding sections. These were done by putting an L meter across the full primary and shorting the individual secondary sections. (L meter uses small signal levels)
1628SE: (with no short = 30 Henry)
4 Ohm Blk to Grn: 50 mH
4 Ohm Blk/Yel to Grn/Yel: 31.4 mH
8 Ohm Blk/Yel to Yel: 43.3 mH
4 to 8 Ohm section Yel to Grn/Yel: 162 mH
Both 4 Ohm sections: 12.6 mH
All sections: 18.8 mH
The rise to 18.8mH from 12.6mH when the 4 to 8 Ohm section is shorted in addition to the two 4 Ohm sections is inexplicable, but sure indicates something is wrong. Also note the rise from 31.4mH to 43.3mH when going from 4 Ohm to 8 Ohm shorted on the same winding. This doesn't happen on the 1640SE below, it falls from 52.8mH to 30mH in that xfmr.
For the 1640SE test readings I have also included a 4x times reading as well. This allows apples to apples comparison with the 1628SE with it 4x higher impedance ratio.
1640SE: (with no short = 7 Henry, 4x = 28 H )
4 Ohm Blk to Grn: 8.2 mH (4x = 32.8 mH)
4 Ohm Blk/Yel to Grn/Yel: 13.2 mH (4x = 52.8 mH)
8 Ohm Blk/Yel to Yel: 7.5 mH (4x = 30 mH)
4 to 8 Ohm section Yel to Grn/Yel: 55 mH (4x = 220 mH)
Both 4 Ohm sections: 4.4 mH (4x = 17.6 mH)
All sections: 4.4 mH (4x = 17.6 mH)
I did some tests on the leakage inductance of the 1628SE and 1640SE versus the secondary winding sections. These were done by putting an L meter across the full primary and shorting the individual secondary sections. (L meter uses small signal levels)
1628SE: (with no short = 30 Henry)
4 Ohm Blk to Grn: 50 mH
4 Ohm Blk/Yel to Grn/Yel: 31.4 mH
8 Ohm Blk/Yel to Yel: 43.3 mH
4 to 8 Ohm section Yel to Grn/Yel: 162 mH
Both 4 Ohm sections: 12.6 mH
All sections: 18.8 mH
The rise to 18.8mH from 12.6mH when the 4 to 8 Ohm section is shorted in addition to the two 4 Ohm sections is inexplicable, but sure indicates something is wrong. Also note the rise from 31.4mH to 43.3mH when going from 4 Ohm to 8 Ohm shorted on the same winding. This doesn't happen on the 1640SE below, it falls from 52.8mH to 30mH in that xfmr.
For the 1640SE test readings I have also included a 4x times reading as well. This allows apples to apples comparison with the 1628SE with it 4x higher impedance ratio.
1640SE: (with no short = 7 Henry, 4x = 28 H )
4 Ohm Blk to Grn: 8.2 mH (4x = 32.8 mH)
4 Ohm Blk/Yel to Grn/Yel: 13.2 mH (4x = 52.8 mH)
8 Ohm Blk/Yel to Yel: 7.5 mH (4x = 30 mH)
4 to 8 Ohm section Yel to Grn/Yel: 55 mH (4x = 220 mH)
Both 4 Ohm sections: 4.4 mH (4x = 17.6 mH)
All sections: 4.4 mH (4x = 17.6 mH)
Hammond 1628 SE output trans
Having just looked at their web to see their connections etc,.
A thought was that if you re measure your frequency response with the secondary connected as 16 Ohm with a dummy 16 Ohm load and it now mesures flatter. Then this, as happens with some output transformers will confirm my belief that there is a design flaw when the whole secondary is not used. The HF resonances is nothing to do with the core gap. that will only effect the low end.
One other point, in your graph when connected as a choke feeding by capacitor another transformer you still have roll off.
This may be because of 'Millar effect' so that the Lundal driving the grid see a lower apparent impedance at HF end. Check with a 10 to 1 scope probe (10pf) monitoring the response at the grid of your 845 tube, to see the flatness of the response to the tube.
Having just looked at their web to see their connections etc,.
A thought was that if you re measure your frequency response with the secondary connected as 16 Ohm with a dummy 16 Ohm load and it now mesures flatter. Then this, as happens with some output transformers will confirm my belief that there is a design flaw when the whole secondary is not used. The HF resonances is nothing to do with the core gap. that will only effect the low end.
One other point, in your graph when connected as a choke feeding by capacitor another transformer you still have roll off.
This may be because of 'Millar effect' so that the Lundal driving the grid see a lower apparent impedance at HF end. Check with a 10 to 1 scope probe (10pf) monitoring the response at the grid of your 845 tube, to see the flatness of the response to the tube.
More leakage inductance tests
I made some more leakage inductance tests on the 1628SE, the 1640SE, and a 1650T P-P xfmr too for comparison, since it has the same secondary configuration.
the 1628SE 4 and 8 Ohm combo winding:
4 Ohm Blk/Yel to Grn/Yel: 31.4 mH
4 Ohm and 8 Ohm Blk/Yel to Grn/Yel and Yel: 42 mH
8 Ohm only Blk/Yel to Yel: 43.3 mH
Notice above that connecting the Yel 8 Ohm lead to the already shorted 4 Ohm leads make the leakage inductance increase from 31.4 mH to 42 mH. This should not be, should go lower or stay the same, not higher.
the 1640SE 4 and 8 Ohm combo winding:
4 Ohm Blk/Yel to Grn/Yel: 13.2 mH (4x = 52.8)
4 Ohm and 8 Ohm Blk/Yel to Grn/Yel and Yel: 7.5 mH (4x = 30)
8 Ohm only Blk/Yel to Yel: 7.5 mH (4x = 30)
The 1640SE behaves normally, with lower leakage with more winding shorted.
Now for the 1650T P-P (1900 Ohm full pri.): (5000/1900 = 2.63)
4 Ohm Blk to Grn: 17.2 mH (x2.63 = 45.2)
4 Ohm Blk/Yel to Grn/Yel: 14.1 mH (x2.63 = 37.1)
8 Ohm Blk/Yel to Yel: 6.7 mH (x2.63 = 17.6)
8 to 4 Ohm Yel to Grn/Yel: 16.4mH (x2.63 = 43.1)
Both 4 Ohm: 4.5 mH (x2.63 = 11.8)
All secondaries shorted: 4.4 mH (x2.63 = 11.6)
and just looking at the 4 and 8 Ohm combo winding of the 1650T:
4 Ohm Blk/Yel to Grn/Yel: 14.1 mH (x2.63 = 37.1)
4 Ohm and 8 Ohm Blk/Yel to Grn/Yel and Yel: 6.2 mH (x2.63= 16.3)
8 Ohm Blk/Yel to Yel: 6.7 mH (x2.63 = 17.6)
Notice that the 1650T behaves normally, always lowering leakage with more shorted windings. Specifically the 8 Ohm combined with 4 Ohm lowers to 6.2x2.63= 16.3 from 14.1x2.63= 37.1 for 4 Ohm only shorted.
I made some more leakage inductance tests on the 1628SE, the 1640SE, and a 1650T P-P xfmr too for comparison, since it has the same secondary configuration.
the 1628SE 4 and 8 Ohm combo winding:
4 Ohm Blk/Yel to Grn/Yel: 31.4 mH
4 Ohm and 8 Ohm Blk/Yel to Grn/Yel and Yel: 42 mH
8 Ohm only Blk/Yel to Yel: 43.3 mH
Notice above that connecting the Yel 8 Ohm lead to the already shorted 4 Ohm leads make the leakage inductance increase from 31.4 mH to 42 mH. This should not be, should go lower or stay the same, not higher.
the 1640SE 4 and 8 Ohm combo winding:
4 Ohm Blk/Yel to Grn/Yel: 13.2 mH (4x = 52.8)
4 Ohm and 8 Ohm Blk/Yel to Grn/Yel and Yel: 7.5 mH (4x = 30)
8 Ohm only Blk/Yel to Yel: 7.5 mH (4x = 30)
The 1640SE behaves normally, with lower leakage with more winding shorted.
Now for the 1650T P-P (1900 Ohm full pri.): (5000/1900 = 2.63)
4 Ohm Blk to Grn: 17.2 mH (x2.63 = 45.2)
4 Ohm Blk/Yel to Grn/Yel: 14.1 mH (x2.63 = 37.1)
8 Ohm Blk/Yel to Yel: 6.7 mH (x2.63 = 17.6)
8 to 4 Ohm Yel to Grn/Yel: 16.4mH (x2.63 = 43.1)
Both 4 Ohm: 4.5 mH (x2.63 = 11.8)
All secondaries shorted: 4.4 mH (x2.63 = 11.6)
and just looking at the 4 and 8 Ohm combo winding of the 1650T:
4 Ohm Blk/Yel to Grn/Yel: 14.1 mH (x2.63 = 37.1)
4 Ohm and 8 Ohm Blk/Yel to Grn/Yel and Yel: 6.2 mH (x2.63= 16.3)
8 Ohm Blk/Yel to Yel: 6.7 mH (x2.63 = 17.6)
Notice that the 1650T behaves normally, always lowering leakage with more shorted windings. Specifically the 8 Ohm combined with 4 Ohm lowers to 6.2x2.63= 16.3 from 14.1x2.63= 37.1 for 4 Ohm only shorted.
For AV8R. Miller effect artifact revealed
Heya. I had to do this measurement anyways, so what better
time than now. I used 12.5 pF probes coz thats all I had.
I measured directly on the 845 grid.
Here are the results and writeup.
Analysis of IT-845-Miller issues with my amp
If you go back and look at my original 845:1628SE plot and
check out the parafeed line, you will see around that
right near 28kHz, where the Miller effect is hitting the 845 grid,
its also slamming reponse seen at the amps output.
-- Jim
Heya. I had to do this measurement anyways, so what better
time than now. I used 12.5 pF probes coz thats all I had.
I measured directly on the 845 grid.
Here are the results and writeup.
Analysis of IT-845-Miller issues with my amp
If you go back and look at my original 845:1628SE plot and
check out the parafeed line, you will see around that
right near 28kHz, where the Miller effect is hitting the 845 grid,
its also slamming reponse seen at the amps output.
-- Jim
Has anyone tried speaking with Hammond?
They seem to be having a lot of trouble catching up with technology. I read a report, about the running of the company, in which their customer service was singled out as a major concern. The report was by Hammond themselves if I remember rightly.
I have emailed them numerous times and never received a reply. I have asked for literature and price lists, and I am still waiting for them to arrive.
If you're trying to sell something, giving your customers the details of your products would be a great place to start. They are a serious annoyance to deal with for me. Very closely followed by Audio Note, who reply to your emails, take your order and then never send it out.
Given the speed with which an email can be sent, that's a relatively worrying service. Especially when it's put aside a company such as RS.
I also read an interesting web site a few weeks ago about the 1628SE. The author tested the OPT's using a square wave generator and points out how one of his OPT's produces over shoots above 10kHz. He ended up testing three or four 1628SE's and noted some quite different characteristics between them all. A few others produced more rounding over, sloped edges, less over shoot and such, but it appears that quality control may be an issue also. Or, at least, consistancy of quality.
They seem to be having a lot of trouble catching up with technology. I read a report, about the running of the company, in which their customer service was singled out as a major concern. The report was by Hammond themselves if I remember rightly.
I have emailed them numerous times and never received a reply. I have asked for literature and price lists, and I am still waiting for them to arrive.
If you're trying to sell something, giving your customers the details of your products would be a great place to start. They are a serious annoyance to deal with for me. Very closely followed by Audio Note, who reply to your emails, take your order and then never send it out.
Given the speed with which an email can be sent, that's a relatively worrying service. Especially when it's put aside a company such as RS.
I also read an interesting web site a few weeks ago about the 1628SE. The author tested the OPT's using a square wave generator and points out how one of his OPT's produces over shoots above 10kHz. He ended up testing three or four 1628SE's and noted some quite different characteristics between them all. A few others produced more rounding over, sloped edges, less over shoot and such, but it appears that quality control may be an issue also. Or, at least, consistancy of quality.
Folks,
I have a pair of the Hammond 1628SE AAMOF. I also measured the dip at about 13-15kHz, but not as great. Just a couple of DB. Then it begins to roll off at about 19kHz, no matter what I do. That is, the frequency response of that transformer is 20-20kHz period. No amount of NFB or tweaking brought anything more out of it.
I used it in a 6L6 and a 300B circuit with similar results.
It is surprising that you are having that severe a roll off problem.
My
Gabe
I have a pair of the Hammond 1628SE AAMOF. I also measured the dip at about 13-15kHz, but not as great. Just a couple of DB. Then it begins to roll off at about 19kHz, no matter what I do. That is, the frequency response of that transformer is 20-20kHz period. No amount of NFB or tweaking brought anything more out of it.
I used it in a 6L6 and a 300B circuit with similar results.
It is surprising that you are having that severe a roll off problem.
My
Gabe
NOT a small dropout
With the output windings configured per standard connection for 8 Ohm output (both 4 Ohm windings in parallel) and an 8 Ohm load on the 8 Ohm points, I see a 55% dropout in voltage (on the 8 Ohm) on the scope at around 15 KHz. Going to higher freq. or lower freq. brings it back up to 100%.
With the output windings standard configured for 4 Ohm output ( both 4 Ohm windings in parallel) and a 4 Ohm load on them, I see a 10% dropout in voltage (on the 4 Ohm) on the scope at around 15 KHz. Going to higher freq. or lower freq. brings it back up to 100%.
Looking at the separate winding voltages in the 8 Ohm load case does not show much more than 10% changes, what is happening is that the 4 to 8 Ohm winding section (Grn/Yel , Yel) is changing phase dramatically around 15 KHz.
This clearly is UNACCEPTABLE performance if you are planning on 8 Ohm output.
With the output windings configured per standard connection for 8 Ohm output (both 4 Ohm windings in parallel) and an 8 Ohm load on the 8 Ohm points, I see a 55% dropout in voltage (on the 8 Ohm) on the scope at around 15 KHz. Going to higher freq. or lower freq. brings it back up to 100%.
With the output windings standard configured for 4 Ohm output ( both 4 Ohm windings in parallel) and a 4 Ohm load on them, I see a 10% dropout in voltage (on the 4 Ohm) on the scope at around 15 KHz. Going to higher freq. or lower freq. brings it back up to 100%.
Looking at the separate winding voltages in the 8 Ohm load case does not show much more than 10% changes, what is happening is that the 4 to 8 Ohm winding section (Grn/Yel , Yel) is changing phase dramatically around 15 KHz.
This clearly is UNACCEPTABLE performance if you are planning on 8 Ohm output.
Smoke,
It is clearly unacceptable at any rate. But... a 50% voltage drop is only a 3dB power drop. (I suspect that saying the 8 ohm config is the two 4 ohm in parallel to be a typo. The two 4 ohm should be in series to give 8 ohms).
So... we can say that the transformer has a 20-15k response at minus 3dB. Not too bad even for hi-fi, IMHO.
I would guess NFB or a capacitor of the right value across the primary would compensate for the phase shift.
Have you or anyone tried at half power, or at 1 watt? Me I test at full power, but still did not get half the voltage at 15 kHz at 8 ohms, with or without NFB. But it did drop.
Maybe I got a fluke pair? I will have to check it out again!
Gabe
It is clearly unacceptable at any rate. But... a 50% voltage drop is only a 3dB power drop. (I suspect that saying the 8 ohm config is the two 4 ohm in parallel to be a typo. The two 4 ohm should be in series to give 8 ohms).
So... we can say that the transformer has a 20-15k response at minus 3dB. Not too bad even for hi-fi, IMHO.
I would guess NFB or a capacitor of the right value across the primary would compensate for the phase shift.
Have you or anyone tried at half power, or at 1 watt? Me I test at full power, but still did not get half the voltage at 15 kHz at 8 ohms, with or without NFB. But it did drop.
Maybe I got a fluke pair? I will have to check it out again!
Gabe
A 55% drop means that 45% voltage is remaining. Power dB is twice voltage dB. (voltage SQUARED, to get power into a constant impedance) A 3dB voltage drop would be a 6 dB power drop. In this case we have a 6.93 dB power drop on the 8 Ohm output.
The 8 Ohm output connection is made by paralleling the two 4 Ohm sections and connecting across the full 8 Ohm winding for output. This is clearly shown on the Hammond box diagram and on their website. Connecting the two 4 Ohm sections in series is how you get the 16 Ohm output, this also is clearly shown on the Hammond wiring diagrams. A simple sanity check here is to observe that twice as many turns (ie. the two 4 Ohm sections) gives 4 times the impedance by the standard calculations for transformers. You cannot just add the individual impedances of the windings.
If your "8" Ohm measurement was performed across the two 4 Ohm sections in series, it would explain why you saw only a small dropout, I only see a 10% dropout for that configuration.
The 8 Ohm output connection is made by paralleling the two 4 Ohm sections and connecting across the full 8 Ohm winding for output. This is clearly shown on the Hammond box diagram and on their website. Connecting the two 4 Ohm sections in series is how you get the 16 Ohm output, this also is clearly shown on the Hammond wiring diagrams. A simple sanity check here is to observe that twice as many turns (ie. the two 4 Ohm sections) gives 4 times the impedance by the standard calculations for transformers. You cannot just add the individual impedances of the windings.
If your "8" Ohm measurement was performed across the two 4 Ohm sections in series, it would explain why you saw only a small dropout, I only see a 10% dropout for that configuration.
smoking-amp said:
Um... I thought that was the opposite. That voltage dB was double power dB. power = 10loga1/a2; voltage = 20loga1/a2. No? Hmmm. When I hook up a pair of speakers in series, they both get the same current but half the voltage... so they each get half power. 3dB is the half power point, no? However, I am probably not considering the liklihood that current also goes down in this case. Sorry. Thinking of other things while writing.
In any event, 3dB is barely noticable with a pure sine wave, and virtually inaudible with music. 6dB is closer to audibly noticing a half amplitude level. Not really a problem if one cannot really hear above 15k anyway, but may affect overall timbre.
Again, though, I agree with you that it is unacceptable. I just have not seen such a drastic dip.
As for Hammond hookup... I was under the impression from what you wrote that the same leads were used for both impedances. For the sake of clarity for those who may not know, the 8 ohm tap is actually another winding in addition to the 4 ohm hooked in parallel. So yes, Smoking-amp is correct, but there are extra windings to make up the 8 ohms, which is in series with the two paralleled 4 ohm hookup.
Again, I will check it again when I get to my lab.
I wonder if that dip and rise of the higher frequencies gives it that "silky" sound?
Gabe
decibels - Oops!
Hi Gabe,
You are absolutely right about the decibels, my mistake.
dB = 20 * log (v1/v2) = 10 * log (power1/power2).
But my dB number was correctly computed at least.
Half voltage gives 6 dB, half power gives 3 dB. For a constant impedance, half voltage gives half current for a 1/4 power result or 6 dB. 0.707 voltage gives half power or 3dB.
Hi Gabe,
You are absolutely right about the decibels, my mistake.
dB = 20 * log (v1/v2) = 10 * log (power1/power2).
But my dB number was correctly computed at least.
Half voltage gives 6 dB, half power gives 3 dB. For a constant impedance, half voltage gives half current for a 1/4 power result or 6 dB. 0.707 voltage gives half power or 3dB.
Heroic effort to fix problem!
I took the end covers off the 1628SE xfmr. to see how it was set up with regards to interleaves. First of all, it is configured differently than the 1640SE. The interconnects to the interleaved layers are different and the 1628SE uses two wires in parallel for the secondary windings while the 1640SE uses a single larger wire for the secondary windings.
Before putting the covers back on, I noticed that the 1628SE had about a 1/16 inch of winding space remaining between the core and laminations. This gave me an idea. I wound another 44 turns on the core with #20 wire. This is the same # of turns as the problematic 4 to 8 Ohm section (Grn/Yel , Yel) ( I confirmed exact voltage matching with a DVM) Then I connected it in parallel (correct phase polarity) with the Grn/Yel, Yel winding. The hope being that this would improve the interleaving effectively. I then re-measured the output voltage versus frequency with an 8 Ohm load connected to the 8 Ohm points. Unfortunately, there was no improvement. A check of the leakage inductance of my added winding then showed it to be poorly coupled (about twice the already poor leakage inductance of the Grn/Yel, Yel winding)
So no surprise it didn't help.
Then I had another idea. If I added another 18 turns for a total of
62 turns and connected my new winding in series with the problematic Grn/Yel, Yel winding section, it would total to the same number of turns as a 4 Ohm winding (106 turns total). Getting 18 more turns into the winding space was a major PITA, but I did get them in. Then I connected the isolated 4 Ohm winding (Blk, Grn) across the combined winding (Blk to Grn/Yel, Yellow to added 62 turns, and the other end of the 62 turns to Grn). This way, the isolated 4 Ohm winding could "enforce" the voltage across the composite 4 Ohm winding, hopefully stiffening the problematic Grn/Yel, Yel section. I then repeated my measurements with an 8 Ohm load across the usual 8 Ohm points. This time I saw an improvement. The voltage that used to dip to 45% at the 15 KHz freq. now it only dips to 60%. Not a big improvement, but since power is voltage squared, this brings the power from .45*.45 = 0.2 up to .6*.6 = .36 (or in dB, a 6.93 dB drop becomes a 4.43 dB drop) I can't say that I recommend this PITA procedure, but what else can one do except try to return the xfmr.
I took the end covers off the 1628SE xfmr. to see how it was set up with regards to interleaves. First of all, it is configured differently than the 1640SE. The interconnects to the interleaved layers are different and the 1628SE uses two wires in parallel for the secondary windings while the 1640SE uses a single larger wire for the secondary windings.
Before putting the covers back on, I noticed that the 1628SE had about a 1/16 inch of winding space remaining between the core and laminations. This gave me an idea. I wound another 44 turns on the core with #20 wire. This is the same # of turns as the problematic 4 to 8 Ohm section (Grn/Yel , Yel) ( I confirmed exact voltage matching with a DVM) Then I connected it in parallel (correct phase polarity) with the Grn/Yel, Yel winding. The hope being that this would improve the interleaving effectively. I then re-measured the output voltage versus frequency with an 8 Ohm load connected to the 8 Ohm points. Unfortunately, there was no improvement. A check of the leakage inductance of my added winding then showed it to be poorly coupled (about twice the already poor leakage inductance of the Grn/Yel, Yel winding)
So no surprise it didn't help.
Then I had another idea. If I added another 18 turns for a total of
62 turns and connected my new winding in series with the problematic Grn/Yel, Yel winding section, it would total to the same number of turns as a 4 Ohm winding (106 turns total). Getting 18 more turns into the winding space was a major PITA, but I did get them in. Then I connected the isolated 4 Ohm winding (Blk, Grn) across the combined winding (Blk to Grn/Yel, Yellow to added 62 turns, and the other end of the 62 turns to Grn). This way, the isolated 4 Ohm winding could "enforce" the voltage across the composite 4 Ohm winding, hopefully stiffening the problematic Grn/Yel, Yel section. I then repeated my measurements with an 8 Ohm load across the usual 8 Ohm points. This time I saw an improvement. The voltage that used to dip to 45% at the 15 KHz freq. now it only dips to 60%. Not a big improvement, but since power is voltage squared, this brings the power from .45*.45 = 0.2 up to .6*.6 = .36 (or in dB, a 6.93 dB drop becomes a 4.43 dB drop) I can't say that I recommend this PITA procedure, but what else can one do except try to return the xfmr.
From the problem to a solution:
Does anyone have any experience with the German transformers from www.askjanfirst.de ?
They have some transformers with similar specs to the Hammond units plus some other (including a 6-8-10k 120ma 20W unit which is very interresting for my 813 project).
Pricing is good, by the way.
Does anyone have any experience with the German transformers from www.askjanfirst.de ?
They have some transformers with similar specs to the Hammond units plus some other (including a 6-8-10k 120ma 20W unit which is very interresting for my 813 project).
Pricing is good, by the way.
I heard from my friend yesterday that the Swedish distributor for Hammond has recieved an email from them where they promissed to check out what´s wrong with the 1628 and if they find anything they will replace the bad units with other ones that meets the specs.
I guess they have to rearrange the windings around the 8 ohms tap or something, then.
I guess they have to rearrange the windings around the 8 ohms tap or something, then.
1628 freq response
I built a pair of 45/2A3 monoblock with Hammond 1628 OPT.
Here's my measurements for 4 and 8-ohm loading with 1V output for mid frequencies:-
10k 11k 12k 13k 14k 15k 16k 17k 18k 19k 20k
8ohm 1.0 .95 .925 .875 .8 .68 .61 .65 .68 .732 .75
4ohm 1.0 .978 .976 .963 .95 .925 .9 .886 .88 .85
Looks like the freq response at 4-ohm loading is much flatter than the 8-ohm. If you don't mind loosing a little more power. Use the 4-ohm tap which would have good frequency response. Also, the distortion will be less too.
Merry Christmas and a happy new year.
Johnny
I built a pair of 45/2A3 monoblock with Hammond 1628 OPT.
Here's my measurements for 4 and 8-ohm loading with 1V output for mid frequencies:-
10k 11k 12k 13k 14k 15k 16k 17k 18k 19k 20k
8ohm 1.0 .95 .925 .875 .8 .68 .61 .65 .68 .732 .75
4ohm 1.0 .978 .976 .963 .95 .925 .9 .886 .88 .85
Looks like the freq response at 4-ohm loading is much flatter than the 8-ohm. If you don't mind loosing a little more power. Use the 4-ohm tap which would have good frequency response. Also, the distortion will be less too.
Merry Christmas and a happy new year.
Johnny
measurements
You need to rewired the secondary wires for getting the 16-ohm output. Therefore, I won't test it unless I could use it. I am just lazy for it.
Regarding measurement in dB's, you can simply use the formula:-
20*log (Vout). - for nominal output of 1V output.
e.g. Vout =0.5V, db = 20*log(0.5) = -6db
Johnny
You need to rewired the secondary wires for getting the 16-ohm output. Therefore, I won't test it unless I could use it. I am just lazy for it.
Regarding measurement in dB's, you can simply use the formula:-
20*log (Vout). - for nominal output of 1V output.
e.g. Vout =0.5V, db = 20*log(0.5) = -6db
Johnny
Reply from Hammond
Here's the message I sent to Hammond and their reply that might interested to you. I will post any reply from Hammond if I got them.
----- Original Message -----
From: Ray Shatzel
To: Johnny
Sent: Thursday, February 24, 2005 8:43 AM
Subject: RE: 1628SE output transformer
Hi Johnny,
THANKS for the info. - we are looking into this - we did get your earlier e-mail but need to verify what is going on here.
We will advise...
Regards
Ray Shatzel
HAMMOND MFG CO
P.S. - In the mean time - if you are unhappy with the items we can gladly refund...but - we need some time in the lab to determine if this is a bad run or a "glitch" that we did not pickup originally...
-----Original Message-----
From: Johnny [mailto:kmtang@direct.ca]
Sent: Wednesday, February 23, 2005 10:08 AM
Subject: Re: 1628SE output transformer
The following test report may be interested to you.
http://www.sacthailand.com/transformerTest1.html
Johnny
----- Original Message -----
From: Johnny
Sent: Wednesday, February 23, 2005 7:00 AM
Subject: 1628SE output transformer
Hi Mark,
I bought a pair of 1628SE output transformer for building of tube amps for myself. I measured the frequency response at 8-ohm ouput and it looks quite disappointing. The response for the 4-ohm is a bit better. There's a sharp dip at approx 14kHz.
I searched in the internet and found the following freq response measurement by Silk Audio in Thailand:- (http://www.sacthailand.com/PIC/TransTestFC-6128SE.jpg). It is rathr close to the measurement I have made.
The specification of the 1628SE transformer from your web site :- Frequency response at least 20 Hz. to 20 Khz. at full rated power (+/- 1 db max. ref. 1 Khz.)
Wondering if you have any comments about this? By the way, the frequency response for the 1627SE is extremely flat that meets the specification.
Have a nice day,
Johnny Tang
Here's the message I sent to Hammond and their reply that might interested to you. I will post any reply from Hammond if I got them.
----- Original Message -----
From: Ray Shatzel
To: Johnny
Sent: Thursday, February 24, 2005 8:43 AM
Subject: RE: 1628SE output transformer
Hi Johnny,
THANKS for the info. - we are looking into this - we did get your earlier e-mail but need to verify what is going on here.
We will advise...
Regards
Ray Shatzel
HAMMOND MFG CO
P.S. - In the mean time - if you are unhappy with the items we can gladly refund...but - we need some time in the lab to determine if this is a bad run or a "glitch" that we did not pickup originally...
-----Original Message-----
From: Johnny [mailto:kmtang@direct.ca]
Sent: Wednesday, February 23, 2005 10:08 AM
Subject: Re: 1628SE output transformer
The following test report may be interested to you.
http://www.sacthailand.com/transformerTest1.html
Johnny
----- Original Message -----
From: Johnny
Sent: Wednesday, February 23, 2005 7:00 AM
Subject: 1628SE output transformer
Hi Mark,
I bought a pair of 1628SE output transformer for building of tube amps for myself. I measured the frequency response at 8-ohm ouput and it looks quite disappointing. The response for the 4-ohm is a bit better. There's a sharp dip at approx 14kHz.
I searched in the internet and found the following freq response measurement by Silk Audio in Thailand:- (http://www.sacthailand.com/PIC/TransTestFC-6128SE.jpg). It is rathr close to the measurement I have made.
The specification of the 1628SE transformer from your web site :- Frequency response at least 20 Hz. to 20 Khz. at full rated power (+/- 1 db max. ref. 1 Khz.)
Wondering if you have any comments about this? By the way, the frequency response for the 1627SE is extremely flat that meets the specification.
Have a nice day,
Johnny Tang
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.