Hi All.
I have recently completed building a pair of horn loaded two way speakers with a split crossover. The crossover is split into HF and LF sections so that one amp is used for the lows and another is used for the highs. The crossover point for this amp is 630 HZ.
My 630 HZ crossover point means that the amp driving the HF section of my speakers doesn't need to be able to put out bass frequencies. Given this fact, what are the optimal characteristics of an OPT for this HF amp?
I build tube amps, so there is no issue with building an amp that has the limitation of working only above 400 HZ (400 HZ rather than 630 HZ because I may want to use it in another 2 way that has a 400 HZ crossover). I am hoping that being freed from the requirement of having to reproduce bass frequencies will create an amp that has better mids and highs than would be the case for a normal full range amp.
The other side of this question is what would be the optimal characteristics of an OPT made for bass frequencies only? Say 400 HZ and below.
My understanding is that the parameters that make OPTs great LF performers are in conflict with the parameters that make them great HF performers. My thought is that my set up allows me to split these opposing requirements into two separate amps.
I know that this is an odd question, but any advice or input would be much appreciated. Also appreciated would be any recommendations for specific OPTs that fit these criteria. Thanks!
I have recently completed building a pair of horn loaded two way speakers with a split crossover. The crossover is split into HF and LF sections so that one amp is used for the lows and another is used for the highs. The crossover point for this amp is 630 HZ.
My 630 HZ crossover point means that the amp driving the HF section of my speakers doesn't need to be able to put out bass frequencies. Given this fact, what are the optimal characteristics of an OPT for this HF amp?
I build tube amps, so there is no issue with building an amp that has the limitation of working only above 400 HZ (400 HZ rather than 630 HZ because I may want to use it in another 2 way that has a 400 HZ crossover). I am hoping that being freed from the requirement of having to reproduce bass frequencies will create an amp that has better mids and highs than would be the case for a normal full range amp.
The other side of this question is what would be the optimal characteristics of an OPT made for bass frequencies only? Say 400 HZ and below.
My understanding is that the parameters that make OPTs great LF performers are in conflict with the parameters that make them great HF performers. My thought is that my set up allows me to split these opposing requirements into two separate amps.
I know that this is an odd question, but any advice or input would be much appreciated. Also appreciated would be any recommendations for specific OPTs that fit these criteria. Thanks!
Interesting question which also applies to my biamping plans.
My ML Aeons crossover to the bass unit at 400HZ and I plan to use a Quad 306 to drive them.
A 40watt PP class A tube amp will drive the electrostatic panels.
I am using Hammond's 1650H OTPs but these may be overkill. Alternatively I could re rate them for the equivalent of 80watts if they only need to cover the 400-20K band.
My thinking is that most amplifier power is required for the low frequencies, which demands a transformer with thicker windings and larger core than would otherwise be required.
So maybe you can just use a smaller OTP eg 10watts assuming the horn speakers are as sensitive as I believe they are?
My ML Aeons crossover to the bass unit at 400HZ and I plan to use a Quad 306 to drive them.
A 40watt PP class A tube amp will drive the electrostatic panels.
I am using Hammond's 1650H OTPs but these may be overkill. Alternatively I could re rate them for the equivalent of 80watts if they only need to cover the 400-20K band.
My thinking is that most amplifier power is required for the low frequencies, which demands a transformer with thicker windings and larger core than would otherwise be required.
So maybe you can just use a smaller OTP eg 10watts assuming the horn speakers are as sensitive as I believe they are?
Last edited:
batteryman: My amps are all single ended, and I currently use a 7-8 watt KT120 on the bass horn, and a 2 watt EL84 amp on the highs. This provides plenty of volume. It seems that single ended OPTs have so many conflicting requirements in order to be be truly full range that using one for HF and one for LF is a near ideal situation. I guess I'm looking for what would be the best commercially available options that already take this into account due to the decisions already made by the OPT designers. However, it seems unlikely that my criteria would be applied given that in most cases the OPT is designed to do as full a range as possible given the design constraints. Alternately, I suppose I could have these made specifically for this purpose by a company that custom makes them, but that is likely to be expensive.
I was going to go SE but the Aeons need >25watts and the load drops to 1.6r at 20Khz so I've had to go with p-p.
I bought a pair of Chinese SE 3.5k OPTs and they are too small compared to a Hammond equivalent so I would expect them being poor at low frequencies but they are probably not going to be good at HF as well.
Sowter say they do a design service at no extra charge so might be worth contacting them.
SINGLE ENDED OUTPUT TRANSFORMERS
I bought a pair of Chinese SE 3.5k OPTs and they are too small compared to a Hammond equivalent so I would expect them being poor at low frequencies but they are probably not going to be good at HF as well.
Sowter say they do a design service at no extra charge so might be worth contacting them.
SINGLE ENDED OUTPUT TRANSFORMERS
Low frequencies require more inductance to avoid saturation (distortion) than high frequencies. More inductance requires larger cores and increased windings. More windings means increased inter-winding capacitance and higher insertion loss. Increased inter winding capacitance affects high-frequency extension. Increased insertion loss makes a transformer less efficient.
If you are using sensitive speakers and have ample power, insertion loss may not be an issue. If one were using a 45 type tube as a SET, the insertion loss is a much bigger deal. Many transformer winders have methods for extending the frequency response of full-bandwidth output transformers and the tradeoff for this is almost always increased cost. The alternative is to use a much smaller transformer with fewer windings. Steps will still need to be taken (ie high pass filter at amplifier input) to prevent a smaller core from saturating as the transformer has no idea it’s bandwidth requirements have been lifted. The high pass filter will also have to be accounted for when implementing a crossover.
George at Tubelab has had a bit of experience swapping output transformers. The best overall (and most expensive) was a full-bandwidth Electra-Print output transformer. The worst was a Hammond. The most efficient was 15 watt $25 Edcor. The Edcor saturated at low frequencies but reputedly had high frequency performance as good as the Electra-Print. I would look at Edcor offerings.
If you are using sensitive speakers and have ample power, insertion loss may not be an issue. If one were using a 45 type tube as a SET, the insertion loss is a much bigger deal. Many transformer winders have methods for extending the frequency response of full-bandwidth output transformers and the tradeoff for this is almost always increased cost. The alternative is to use a much smaller transformer with fewer windings. Steps will still need to be taken (ie high pass filter at amplifier input) to prevent a smaller core from saturating as the transformer has no idea it’s bandwidth requirements have been lifted. The high pass filter will also have to be accounted for when implementing a crossover.
George at Tubelab has had a bit of experience swapping output transformers. The best overall (and most expensive) was a full-bandwidth Electra-Print output transformer. The worst was a Hammond. The most efficient was 15 watt $25 Edcor. The Edcor saturated at low frequencies but reputedly had high frequency performance as good as the Electra-Print. I would look at Edcor offerings.
You're welcome tizman, and yes Brinkman outlines what I was thinking.
While some OPTs need to resort to fancy winding schemes to get higher frequency performance, it's interesting what you'll find naturally on a smaller unit which has poorer lowest frequency performance. I have been using a stereo pair, of 1960's Japanese origin from lower power amps. The higher frequencies are not an issue.
I happen to have sensitive speakers FWIW, I need much less than a watt unless I take it closer to 200Hz, and yes, I'm careful to keep anything lower out of it.
While some OPTs need to resort to fancy winding schemes to get higher frequency performance, it's interesting what you'll find naturally on a smaller unit which has poorer lowest frequency performance. I have been using a stereo pair, of 1960's Japanese origin from lower power amps. The higher frequencies are not an issue.
I happen to have sensitive speakers FWIW, I need much less than a watt unless I take it closer to 200Hz, and yes, I'm careful to keep anything lower out of it.
Brinkman: Thanks for the explanation. The way I have the system set up now, both amplifiers get the full signal, and then each one drives its own section of the crossover. I hadn't thought about the implications of providing a full signal to both amps and thereby the amps' OPTs with respect to saturation. Perhaps it would be useful to install a passive line level crossover on the input of each amplifier to (largely) remove the frequencies of the signal that the amplifier will not be required to reproduce when driving their respective portions of the crossover and speaker. If my crossover is at 630 HZ, I can put together two passive line level crossovers that operate on the inputs of the HF and on the input of the LF amps so that the amplifiers do not receive the majority of the signal they are not required to reproduce. Does this make sense?
Great info. I have many small OPTs that might be appropriate for the upper frequencies, and a few large OPTs that have great bass but poor HF quality. Brinkman makes a good point about the signal going to both amps either way, and that that is also an issue. It's easy to add PLLs that are designed to eliminate that problem to the inputs of the amps. Perhaps passive line level crossovers are the way to go rather than the regular passive speaker crossovers i have at the moment.
Yes, you can have someone design and make transformers that are optimized for the upper or lower part of the audio spectrum. Do you think:
1. That would result in lower cost for you compared to buying full-spectrum off the shelf OPTs?
or
2. That you need OPTs of a quality/performance level unavailable off the shelf?
I have a lot of parts at home, and ideally would like to use what I already have at hand. My goal is to use what I have and if the weaknesses in the parts I already have are avoided by using them this way, or if those weaknesses amount to strengths because of my specific application, all the better. To answer your questions...
1. I really doubt that I could have two purpose built custom transformers made for less than a single, quality, off the shelf unit.
2. The quality and performance level that I would like to attain are specific to the particular use I have in mind. I think a competent OPT manufacturer could easily design and build the two separate OPTs that I would use in my system that would outperform a single OPT used normally. I believe this is the case because of the difficulty inherent in building a single full range high quality OPT.
1. I really doubt that I could have two purpose built custom transformers made for less than a single, quality, off the shelf unit.
2. The quality and performance level that I would like to attain are specific to the particular use I have in mind. I think a competent OPT manufacturer could easily design and build the two separate OPTs that I would use in my system that would outperform a single OPT used normally. I believe this is the case because of the difficulty inherent in building a single full range high quality OPT.
Your input coupling capacitor is a freebie PLL, I like that kind of creativity.
In cases like this it can be worth reverse engineering your passive speaker level crossover to put the filters before or within the amp. This leaves speaker EQ, which you can leave at speaker level or do before the amp, incorporating extra or needed rolloff.. or within the amp if you want.
In cases like this it can be worth reverse engineering your passive speaker level crossover to put the filters before or within the amp. This leaves speaker EQ, which you can leave at speaker level or do before the amp, incorporating extra or needed rolloff.. or within the amp if you want.
My amps are all two stage with a single dual triode driver tube followed by a pair of output tubes. All my amps have a 100K volume pot at the input, and some have a small DC blocking capacitor before the volume pot. Do you mean that in the case of the PLLXO high pass, the input capacitor that I sometimes include before my volume pot can be part of the PLLXO high pass section?
To your first question, you can use any combination of speaker level and line level filter, however at this time the focus is on relieving the amp from the bass.
To your second question, yes, why not use an existing capacitor like that. It saves you trying to buffer impedances and makes for one less thing to go wrong.
BTW, my amp is broadly similar to yours.
To your second question, yes, why not use an existing capacitor like that. It saves you trying to buffer impedances and makes for one less thing to go wrong.
BTW, my amp is broadly similar to yours.
Alternatively, one could use a quality subwoofer plate amplifier with internal DSP. Many of them have line level outputs that can be connected to your tube amplifier and the DSP can be implemented to set whatever high pass frequency you desire, at whatever slope works best for your setup. This is assuming you would also be using the plate amplifier to drive a dedicated subwoofer or midbass driver.
I am in a very similar situation with a Tubelab single ended amplifier using a 45 tube and the aforementioned 15 watt Edcor output transformer. The above scheme is how I am going to sidestep the output transformer saturation concern and the midbass driver, low passed at 200Hz, will eliminate loudspeaker floor-bounce and baffle step compensation from the mains.
I am in a very similar situation with a Tubelab single ended amplifier using a 45 tube and the aforementioned 15 watt Edcor output transformer. The above scheme is how I am going to sidestep the output transformer saturation concern and the midbass driver, low passed at 200Hz, will eliminate loudspeaker floor-bounce and baffle step compensation from the mains.
A HF transformer can be wound on a small high quality core, like 80% Ni, with primary inductance of 5÷10 H. Don't have to bother with winding interleaving because core coupling will dominate even at highest audio frequencies. Leakage inductance and stray capacitance will resonate in far supersonic area, so won't be a problem, too.
For LF transformer, only inductance will matter, and it can be easily made very high because parasitics won't matter.
Dividing the audio spectrum into three rather than two bands will make transformer design even easier.
For LF transformer, only inductance will matter, and it can be easily made very high because parasitics won't matter.
Dividing the audio spectrum into three rather than two bands will make transformer design even easier.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.