Pano,
You may be onto something.
The data sheet for the 10K : 16 version states that the DC resistance is 383 ohms.
If it turns out that you do have a 5K : 8 version perhaps you can report the DC Resistance of the primary. That would be a clue.
DT
Scripture refrence please
What translation, chapter, and verse are you reading? RDHB 4th edition? The 4th Edition is what is on my shelf.
BTW: I washed my Dad’s puck-up truck a lot of times when I was a kid for that book!
DT
What translation, chapter, and verse are you reading? RDHB 4th edition? The 4th Edition is what is on my shelf.
BTW: I washed my Dad’s puck-up truck a lot of times when I was a kid for that book!
DT
Did anyone calculate the actual loss for various source impedance i.e. plate resistance?
I tried to validate the Edcor claim and found to my surprise that 25H is indeed enough inductance with 5K reflected load in parallel to be less than -1dB at 40 Hz, relative to the midband level, for a range of source impedance up to 2K ohms. So for reasonable damping factors of 2+ the claim seems to bear out. I didn't do a spice model but a simple bounding analysis is sufficient to show.
Those values are for pentodes without feedback, i.e. open loop. To which feedback would be applied to reduce the open loop gain and extend the closed loop frequency response.
If one assumes a triode with some reasonable plate resistance say 1200 ohms for the 5K load and 2400 ohms for the 10K load, the required inductance for Ra = 968 ohms and 1936 ohms are respectively about 8.5H and 17H.
The highest plate resistance that can be used with the 10K reflected load and 25K Lpri for -1 dB at 40Hz appears to be around 5K ohms, or 801A territory.
So Edcor's claim does have a basis given some reasonable assumptions.
It still doesn't answer the original question but refutes the obviousness of the 10K unit being a relabeled 5K unit.
Cheers
Agreed, Lowering the Rp, or proper negative feedback to reduce Rp, thus increasing the low frequency response.
Agreed, one can reduce Ra and easily reach -1db at 20hz or lower.
Yes, by lowering Ra using a variety of triodes or using negative feedback.
Cheers.
Ra includes the load resistance referred to the primary, Rp of the tube, and dc wire resistance. For pentode Ra is approximately the load resistance referred to the primary, which is 5k. Sounds like your model has a 5k resistor in parallel with the transformer primary, thus approximately 2.5k, not 5k. With such 25 henries satisfies
-1db at 40hz.
Would you mind posting the model you used?
Cheers.
Last edited:
Refuted? Not yet!
I do not recall anyone stating that the 10K unit was a relabeled 5K, only that it is possible that they are the same. Likely? We do not know yet!
While you guys are discussing the applicable math and the low frequency extension of the transformers you have neglected the High frequency end of the bandwidth.
From the discussion so far the primary inductance of the 10K unit will meet the performance specification at the LF end of the bandwidth for both applications. The HF performance has not been discussed.
The theory and math is all lined out in chapter 14 of the 1957 Electronic Designers Handbook by Landee, Davis Albrecht.
DT
All I said is that 25H is not obviously wrong for 10K. It depends on one's assumptions. All I assumed is a reasonable plate resistance in parallel with the OPT. No one uses open loop pentode amps.
I'm making no other claims about these transformers as there is no published data. Who knows about the high end. Inductance has nothing to do with the high end.
I do have some trust developed with Edcor as they have supplied good product to me in the past. Maybe they know what they're doing. What seems silly to me is to speculate on zero information besides the ratio that they are the same unit simply relabeled.
Last edited:
I used the same model used to make the table in RDH4 P.213 which is just the voltage divider equation using vector arithmetic.
Using the RDH table, the equivalent impedance value to use is 5K in parallel with 2K plate resistance worst case (open loop pentode amps are very rare) for an effective Ra of about 1400 ohms. This gives 16H at 30 Hz from the table.
All I'm saying is that 25H is enough inductance to build a practical tube stage with a triode (or by extension a pentode/ultralinear with some NFB) and 5K or even 10K reflected impedance.
I did not originally intend for this aspect to be entered into the discussion, just that the two transformers are different. However, if one checks, 25 henries only has approximately 6.3k ohms reactance at 40hz.
RCA RDH4 says more is needed.
Here is a post from Johan Potgieter, which is also basically agrees with RDH4 and me.
http://www.diyaudio.com/forums/tubes-valves/157920-choosing-output-transformer.html
We are dealing with 6.3k from 25H with Ra of 5k and 10k As one can see, -1db is impossible. But as Michael mentions, using a triode makes the point somewhat mute. Using feedback also helps.
Further correct information provided by Johan
Ra in this case is approximately 3.62k. However we are dealing with Ra of 5k and 10k. Let's change the conditions Johan has, one at a time.
So for Ra = 3.62k, we need 28.6hy for -3db at 20hz.
Doubling the frequency to 40hz we need 14.3hy for -3db.
For Ra of 5k instead of 3.62k we need 19.75hy for -3db.
For -1db, Ra =5k, we need to multiply the inductance by approximately 2.76, for approximately 39.4hy for -1db at 40hz.
For 10k we need approximately 78.8hy for -1db at 40hz.
Fairly close to 52hy and 87hy RHD4 lists. Of course we rolled off, not counted wiring resistance, loses etc. As I stated earlier, there is probably just a misprint somewhere on the spec sheet.
Cheers.
Last edited:
I most heartedly agree Michael. You are absolutely correct that the plate Rp makes the design work nicely.
My original point was that the two transformers were different as per the OP question. I guess some saw the RHD4 numbers, which are correct, and that there was a discrepency problem. I had no such intentions of bringing that aspect into the conversation, just that the two transformers were different. And there is probably just a typo on the page.
Cheers Michael and thanks for your insights.
Last edited:
No typos or misunderstanding as far as I can tell. And it is important to the core discussion here to understand that it's just the matter of assuming a practical value for the plate resistance of the triode. The Ra figure in RDH4 includes plate resistance in parallel. When you add this in you get Ra values on the order of 2K or less and the required inductance is in range of the Edcor unit.
Negative feedback with a pentode accomplishes the same result.
My point is that it is an unrealistic and non-useful assumption to use the open loop pentode Ra figure to determine if a given transformer is suitable based on the -1 dB criteria. No one will build and use an open loop pentode amp.
If someone eventually comes up with inductance and DCR figures for both transformers then this will still be a useful discussion.
Cheers
To reply to your last statement, the transformers are different according to the specs given, no doubt about that. That question has been answered along time ago.
RDH4's table lists Ra from 800 to 50,000 ohms.
To avoid confusion, my typo comment refers to the manufacturer's printed specs, not your comment. In fact, the manufacturer did list the specs using 5k and 10k, did they not? As such, the given specs are not correct, period. It is good to see that you are thinking of your friends and am helping them to improve the bass response with the transformers.
If the inductance were more, the proper value to give the -1db spec, then the bass would be more and deeper. Adding a triode or feedback to lower the Ra would add even more bass as a result. It certainly would require alot less feedback in pentode or even triode operation. (By the way, one does not need feedback in pentode operation. And especially global feedback has its drawbacks, but that is for another string.)
And using a pentode with feedback does not necessarily give one the correct bass response since the feedback could easily be less than required. The amount will depend upon the particular design, open loop gain etc.
The fact is we can improve the transformers performance, the systems performance, there is no doubt. However, the specs, as given by the manufacturer, are not correct.
Cheers.
ps. But for $37, who is looking for, or expecting, the worlds best transformer?
Last edited:
I returned some GXSE 15-8-5K OTs for GXSE 15-16-10K ones. They are definitely NOT the same OT. The The 8 Ohm sec has 100 Turns and the 16 Ohm secondary has 145 Turns.
GXSE 15-8-5K:
Primary DC Ohms: 159
Primary inductance (no DC): 13.2 H on L/C meter, 19 H at 15 VAC 60 Hz excitation (at secondary), and 83 H with the gap spacer removed (pressure clamped) at 13.5 VAC 60 Hz (at secondary).
Maximum linear excitation 24 VAC at 60 Hz on 8 Ohm sec before magnetizing current spiking occurred.
GXSE 15-16-10K:
Primary DC Ohms: 375
Primary inductance (no DC): 26.4 H on L/C meter (as also expected by turns calc.) and 38 H with
21 VAC 60 Hz excitation (at secondary)
Both OTs have the same thickness gap spacer. Core area is 1" by 1.25".
DC current must, at a minimum, further reduce the inductance ratings by 1/2 due to the DC flux using half the rated core flux. So reduce the above L#s by at least 1/2 for SE operation. (so 5K primary is less than 10 Henry with DC present)
Calculating power rating (using the 8 Ohm data): 24 VAC at 60 Hz gets reduced to 16 VAC at 40 Hz. Then due to max DC using 1/2 of the core flux, this gets reduced to 8 VAC at 40 Hz with DC present. Giving 8x8/8 = 8 Watts max capability. Clearly the given Watt ratings are overly optimistic, since a SE GXSE-15 OT is of the same size as the GXPP-15, with the same LF end, so it must give 1/4 the power due to the DC flux issue for the same max core flux.
(so really only 4 SE Watts at comparable saturation to the P-P OT if the same number of turns are used. The GXPP actually uses more turns.)
I was quite dis-satisfied with the GXSE offerings both power and freq.-wise and so added a DC buck winding onto the remaining core window space, plus I used 2X the rated impedance rating to get it to work properly for freq. for a 6BQ5 SE amp. (The 10K primary will have .707 the max DC current capability of the 5K primary due to 1.4X turns.)
Just for comparison, a CXPP-25 P-P OT draws 1/48 the magnetizing current of the GXSE-15 at max power rating (25 Watt versus 8 Watt too), cost $10 more, and has a 20 Hz to 20 KHz rating. Of course you have to use an anti-triode circuit mode to use the P-P OT.
Never bought any more SE OTs of any kind since. Total crap, if you bother to actually measure them.
GXSE 15-8-5K:
Primary DC Ohms: 159
Primary inductance (no DC): 13.2 H on L/C meter, 19 H at 15 VAC 60 Hz excitation (at secondary), and 83 H with the gap spacer removed (pressure clamped) at 13.5 VAC 60 Hz (at secondary).
Maximum linear excitation 24 VAC at 60 Hz on 8 Ohm sec before magnetizing current spiking occurred.
GXSE 15-16-10K:
Primary DC Ohms: 375
Primary inductance (no DC): 26.4 H on L/C meter (as also expected by turns calc.) and 38 H with
21 VAC 60 Hz excitation (at secondary)
Both OTs have the same thickness gap spacer. Core area is 1" by 1.25".
DC current must, at a minimum, further reduce the inductance ratings by 1/2 due to the DC flux using half the rated core flux. So reduce the above L#s by at least 1/2 for SE operation. (so 5K primary is less than 10 Henry with DC present)
Calculating power rating (using the 8 Ohm data): 24 VAC at 60 Hz gets reduced to 16 VAC at 40 Hz. Then due to max DC using 1/2 of the core flux, this gets reduced to 8 VAC at 40 Hz with DC present. Giving 8x8/8 = 8 Watts max capability. Clearly the given Watt ratings are overly optimistic, since a SE GXSE-15 OT is of the same size as the GXPP-15, with the same LF end, so it must give 1/4 the power due to the DC flux issue for the same max core flux.
(so really only 4 SE Watts at comparable saturation to the P-P OT if the same number of turns are used. The GXPP actually uses more turns.)
I was quite dis-satisfied with the GXSE offerings both power and freq.-wise and so added a DC buck winding onto the remaining core window space, plus I used 2X the rated impedance rating to get it to work properly for freq. for a 6BQ5 SE amp. (The 10K primary will have .707 the max DC current capability of the 5K primary due to 1.4X turns.)
Just for comparison, a CXPP-25 P-P OT draws 1/48 the magnetizing current of the GXSE-15 at max power rating (25 Watt versus 8 Watt too), cost $10 more, and has a 20 Hz to 20 KHz rating. Of course you have to use an anti-triode circuit mode to use the P-P OT.
Never bought any more SE OTs of any kind since. Total crap, if you bother to actually measure them.
Last edited:
correction:
The GXPP-15 uses LESS turns than the GXSE-15, since the core is not gapped or DC encumbered. Uses 62 Turns versus 100 Turns for an 8 Ohm sec.
By the way, it is easy to measure the actual turns on any OT, just wind 10 turns of wire wrap wire into the remaining winding window and measure the AC voltage on it and on the primary when the secondary is driven at low VAC 60 Hz. (be carefull of the High voltage present on the primary, connect clip leads before powering up, and do not use too much voltage on the secondary or it will exceed the safety V breakdown rating of the primary. Just a few volts is usually sufficient on the secondary if you have a 5 digit DVM to measure the 10 turn winding V.) You must separate the measured windings from the driven windings to get an accurate measurement, otherwise DC resistance in the driven winding corrupts the reading.
The GXPP-15 uses LESS turns than the GXSE-15, since the core is not gapped or DC encumbered. Uses 62 Turns versus 100 Turns for an 8 Ohm sec.
By the way, it is easy to measure the actual turns on any OT, just wind 10 turns of wire wrap wire into the remaining winding window and measure the AC voltage on it and on the primary when the secondary is driven at low VAC 60 Hz. (be carefull of the High voltage present on the primary, connect clip leads before powering up, and do not use too much voltage on the secondary or it will exceed the safety V breakdown rating of the primary. Just a few volts is usually sufficient on the secondary if you have a 5 digit DVM to measure the 10 turn winding V.) You must separate the measured windings from the driven windings to get an accurate measurement, otherwise DC resistance in the driven winding corrupts the reading.
Last edited:
OK, I see the confusion. The manufacturer's printed specs are ambiguous as to the load vs the source impedance. BTW they aren't my "friends" and I'm a little offended that you are suggesting some sort of bias on my part. I'm making a pure practicality argument here.
I am assuming that specifying a transformer for 5K reflected load or 10K reflected load is making underlying assumptions that it will be used in a practical amplifier (triode or NFB pentode) with the specified reflected load. As such the -1 dB at 40 Hz point will be met in practice.
And of course increasing the inductance will improve the bass response. And you can pay more money and get that. Even with Edcor. My only claim is that the Edcor GXSE15 transformers actually do have enough inductance to meet their frequency response spec. in practical amplifiers at their rated reflected load.
I am assuming that specifying a transformer for 5K reflected load or 10K reflected load is making underlying assumptions that it will be used in a practical amplifier (triode or NFB pentode) with the specified reflected load. As such the -1 dB at 40 Hz point will be met in practice.
And of course increasing the inductance will improve the bass response. And you can pay more money and get that. Even with Edcor. My only claim is that the Edcor GXSE15 transformers actually do have enough inductance to meet their frequency response spec. in practical amplifiers at their rated reflected load.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.