Curious what test configuration was used and what were the results? I've got a pair of the 40W 4.8k in my closet that I was hoping to use some day, but you got me wondering if it's even worth a try.
The top end looks rolled off by the manufacture supplied curves, but the measurements were probably taken using a source with a Zout = the reflected impedance. I'd suspect that acceptable bandwidth may be attainable in circuit if driven by triodes with plate impedance in the 1/2-1/3rd range (typical triode loading) of the transformers primary Z. And/Or with use of NFB.
Last edited:
I talked to LeFevre about this awhile ago. His call was that, though toroids look cool to people due to the optimized iron circuit configuration, they cannot be made to have fine resolution due to the severely compromized copper circuit config. The winding and interleaving lay is very screwed up by the huge circumference difference in the outside and inside of the core. On the inside, the windings are piling up; on the outside, the windings are spaced apart, and the next (interleaved) layer is falling into the spaces.
In practice, this appears to be true. There is an earlier product designed by Van der Veen, which I have not tried either. I haven't heard much about them.
Aloha,
Poinz
AudioTropic
In practice, this appears to be true. There is an earlier product designed by Van der Veen, which I have not tried either. I haven't heard much about them.
Aloha,
Poinz
AudioTropic
Actually pretty widely used, and properly designed toroid OPTs are fine as long as you make sure there is no significant level of dc flux in the core.
Both Amplimo and Plitron make van der Veen designed toroidal OPTs.
Menno van der Veen, audio electronic research & consultancy
FWIW I think the winding argument stated in an above post is a bit specious - toroids are rather commonly used in non-audio specific applications where very wide bandwidth and good linearity are required. One can loosely make the same argument for the EI transformer, and what is often the case is that a toroid may offer much higher permeability in a given volume and much lower excitation losses as well. The big problem is it is hard to make toroids that can handle significant dc flux, I have seen this done with cut cores, but have not worked with or heard an amplifier based on such a core.
Last edited:
I have advocated and used power toroids in almost all of my amps as OT. They have always worked well with bandwidth from 10hz to at least 50khz. The theoretical critisms of toroidals as outputs just don't map to my experience. They win over EI on having clean bandwidth, and no hysterics losses which would tend to blur low level detail at crossover and low signal levels (ie almost all of the time).
This is for severely none optimized toroidals. The Piltron ones are highly optimized and are supposedly the most neutral OT available. The Antek ones seem to be very badly designed, they probably tried interleaving, which boosts up the already high interwinding capacitance.
My toroidals will tolerate little more than 1mA of imbalance and because of this I use carefully matched CCS (simple LM317 type) to impose DC balance. Alternatively Garter bias will balances things adequately.
Shoog
This is for severely none optimized toroidals. The Piltron ones are highly optimized and are supposedly the most neutral OT available. The Antek ones seem to be very badly designed, they probably tried interleaving, which boosts up the already high interwinding capacitance.
My toroidals will tolerate little more than 1mA of imbalance and because of this I use carefully matched CCS (simple LM317 type) to impose DC balance. Alternatively Garter bias will balances things adequately.
Shoog
Thanks Shoog, can you point me in a direction to learn how to use CCSs to balance the DC? I have some Anteks to use and would like to see how it goes before bailing and buying othe OPTs.
I got a pair of "400 watt at 20Hz" Plitrons from their surplus page a few years ago. I have had them to 650 volts and 250 watts which was idling for them. The frequency response at 250 watts was below 10Hz (measurement limit) to 65 KHz. They begin to show signs of saturation at 200 W and 20 Hz when the DC offset goes over 25 mA or so. They are indeed very good sounding transformers, although some have critisized them for being too clean. Of course purchased new they are real expensive.
I have tried several power toroids as OPT's with varying results. Some just refuse to work Some work OK and a few work great. The best that I found were some dumpster toroids from the trash behind a shop that fixed thermal printers. They appear to ba about 200 VA. They are tolerant to about 5 mA of imbalance and work to 50 watts at 20 Hz. I have run them to 450 volts of B+.
Try this for size, its where I got the idea;
The Tube CAD Journal,Three-Pin Regulators
The power toroidals I have used are all the same manufacturer - European unfortunately - I think the brand is Telema. I found the polarity of the primary quite critical as to induced resonances - so try both directions. The winding stratergy is quite straightforward been bifilar primary next to the core and then biflar secondary on top. No screening, nothing.
Shoog
The Tube CAD Journal,Three-Pin Regulators
The power toroidals I have used are all the same manufacturer - European unfortunately - I think the brand is Telema. I found the polarity of the primary quite critical as to induced resonances - so try both directions. The winding stratergy is quite straightforward been bifilar primary next to the core and then biflar secondary on top. No screening, nothing.
Shoog
I used them unpotted at about 160VA for 7W of power.
I have generally found it necessary to unwind the secondary to get the ratio as the batch I had were 18V and my load is always 4ohms. Sometimes I have just wound a whole new secondary over the top and simply ignored the prewound secondary. Very easy to get whatever ratio you need that way. Theoretically you could wind a copper screen on as well.
Shoog
I have generally found it necessary to unwind the secondary to get the ratio as the batch I had were 18V and my load is always 4ohms. Sometimes I have just wound a whole new secondary over the top and simply ignored the prewound secondary. Very easy to get whatever ratio you need that way. Theoretically you could wind a copper screen on as well.
Shoog
@Jeb-D
Check out this thread: http://www.diyaudio.com/forums/tubes-valves/144443-experience-antek-opts.html
I did a few tests. Nothing that could be used to make a datasheet, but the results were consistent. Huge drop off after 6k. It is apparent there was no quality control.
Check out this thread: http://www.diyaudio.com/forums/tubes-valves/144443-experience-antek-opts.html
I did a few tests. Nothing that could be used to make a datasheet, but the results were consistent. Huge drop off after 6k. It is apparent there was no quality control.
Ouch!
I'll defiantly have to do a measurement on mine prior to building an amp around them.
I have near zero experience with either Edcor or Hammond for P-P OPT's. I will likely order a pair of the Edcor 100 watt P-P OPT's when the budget allows. I bought some of the bottom of the line Edcor and Hammond SE OPT's and tested them. XSE15-8-5K and 125 CSE. I also tried the top of the line 1628SE, 1628SEA, CSXE25-8-5K.
In the small size the two brands tested and sounded almost the same.
In the large size the Hammond 1628SE was terrible with a severe notch of 10 to 20 db at about 18KHz resulting in a HF rolloff starting at 10 KHz. It has been replaced by the 1628SEA. The 1628SEA is much better but still has a 3 db loss at 20 KHz. The Edcor CXSE gives up a tiny bit of bass capability below 20 Hz for better HF response, but is still 1 to 2 db down at 20 KHz. The Edcor sounds more alive than the Hammond and has better efficiency so the amp makes about a watt more power. The differences between the two transformers are magnified with high efficiency speakers where tha average power is below a watt. On 106db Lowthers the big Hammond lacks detail. The Edcor was not tried but might have done the same. This verifies that an OPT can be too big, the big core just takes too much energy to magnetize.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.