4/8 ohm speakers on single tap OPT

[tbobbyt]

Ok, I’m trying to get started on building a Simple SE amp. I have the SE board, a matched pair of 6L6GC tubes and an old pair of KLH 20 4 ohm speakers that I would like to use for now out of necessity. Another set of 8 ohm speakers may be added sometime down the road, but not at this time. I am trying to choose the other components.

So I would like to use the 4 ohms for now, but be able to also use 8 ohms later. I like the One Electron UBT-2’s with the multiple taps, but can’t afford the price of them at this time. The ones that I can afford are the Edcor GXSE15’s with the 5K ohm load. However I have to choose between the GXSE15-8-5K and GXSE15-4-5K.

From what I understand with the 4-5K’s I’d never be able to run 8 ohm speakers in the future, 8-10K is no good, but with the 8-5K’s I may have a chance of using the 4 ohm speakers even with the load dropping to 2.5K.

I don’t have a PT yet, but I have read that a 375-0-375 results in a B+ of 440-460 volts, and high B+ requires fairly high load impedance. So I’m thinking of using a power supply that will give a lower B+ to require less load impedance. Is that right? I’m not sure where a B+ sweet spot may be to use with the 4 ohm 2.5K load now and the 8 ohm 5K later. Then again I’m not sure how much the B+ would drop if I use a PT with 350-0-350 or 360-0-360 and how much the load impedance requirements would change. Is this even a good idea to try, or should I wait till I can afford the UBT-2’s?

Any ideas on this would be appreciated, along with any other thoughts on using 4/8 ohm speakers without an output tranny that has multiple taps.

Thanks,
Bob

[rknize]

Indeed, that is a limitation of the Edcor SE series. I would give them a call and ask if they can add a tap for you. Otherwise the UBT-2's are a fine choice, but notice they lack an UL tap on the primary side. Hammond is an option, though not the most ideal. Is Transcendar iron still available? There are other more pricey options too, like James and so forth.

[tomchr]

Originally Posted by tbobbyt

Ok, I’m trying to get started on building a Simple SE amp.

Sweet! I've built a couple of amps based/inspired by the SSE. It's a good design. I liked it with my 300B's...

Originally Posted by tbobbyt

I like the One Electron UBT-2’s with the multiple taps, but can’t afford the price of them at this time. The ones that I can afford are the Edcor GXSE15’s with the 5K ohm load. However I have to choose between the GXSE15-8-5K and GXSE15-4-5K.

Just keep in mind that with OPTs you get what you pay for. The One Electron ones are supposed to be quite good. As are the Edcor CXSE25-series. The ElectraPrint transformers are wound to your specs and are supposed to be a notch better still (but easily run $200/EA for versions with 4 and 8 ohm secondaries).

Originally Posted by tbobbyt

From what I understand with the 4-5K’s I’d never be able to run 8 ohm speakers in the future, 8-10K is no good, but with the 8-5K’s I may have a chance of using the 4 ohm speakers even with the load dropping to 2.5K.

If I understand you correctly, then you're right. A transformer transforms the impedance from the secondary to the primary (and vice versa). There's no magic to it. An OPT designed for 5 kOhm primary with 8 ohm secondary load will present 2.5 kOhm on the primary when loaded by 4 ohm on the secondary. Some amps/tubes may be fine with this change in load impedance, others may not. Whether that's the case for the 6L6 or not, I can't tell you as I have no experience with that tube. I'm sure others will chime in though.

Originally Posted by tbobbyt

I don’t have a PT yet, but I have read that a 375-0-375 results in a B+ of 440-460 volts, and high B+ requires fairly high load impedance. So I’m thinking of using a power supply that will give a lower B+ to require less load impedance. Is that right?

Theoretically, you should get sqrt(2)*Vrms across the reservoir cap. So 1.414*375 = 530 V in case of a 375-0-375 PT. However, that requires that the rectifiers conduct an infinite amount of current for an incredibly short amount of time every charging cycle (zero conduction angle). In practice, you can expect close to that voltage with zero load. At load you'll see about 1.2*Vrms or 1.2*375 = 450 V.
Above calculations don't take the drop across the rectifier into account. For silicon diode rectifiers, the drop can be neglected for HV supplies. But for tube rectifiers, you need to take the voltage drop into account. It depends on load current (see the datasheet for the rectifier tube) and is generally on the order of 50 V or so. You need to subtract that from the voltage calculated above.

Some people swear by a power supply design software called PSUD. Others - like myself - just draw the equivalent circuit diagram including transformer DC resistances and do the math. Or better yet.... use a PT from something else and crank the variac until the amp works right...

Originally Posted by tbobbyt

I’m not sure where a B+ sweet spot may be to use with the 4 ohm 2.5K load now and the 8 ohm 5K later.

I don't know either (no experience with that tube). But unless the difference in B+ voltage is extreme between the sweet spots of 2.5 kOhm and 5 kOhm loads, you could just design for the higher B+ and drop any excess voltage across a resistor if you end up needing a lower voltage. Just beware of the power dissipated in this dropper resistor... As I recall the SSE already has some resistance in the supply as part of the RC filter. 150 ohm, 5 W or something. Just change that until the B+ is right. You'll need to measure B+ at load for this.

Originally Posted by tbobbyt

Is this even a good idea to try, or should I wait till I can afford the UBT-2’s?

Depends on your goals. Personally, I'd wait. But if you'd rather get started now, by all means do so. Just beware that you'd probably end up spending more by having to re-tweak the design later.

Originally Posted by tbobbyt

Any ideas on this would be appreciated, along with any other thoughts on using 4/8 ohm speakers without an output tranny that has multiple taps.

You could try to find a transformer that has a 6 ohm secondary. That's a cost-effective compromise. But that probably sends you off towards a custom wound transformer.

You could also use two 6L6'es in parallel for the 4 ohm configuration. If one 6L6 requires a 5 kOhm primary load, I would think that two in parallel would want a 2.5 kOhm load... Options, options.

~Tom

[spopepro]

I think the most sensible thing to do is to wait until you can buy the big iron. It has been stated many times that the the OPT is such a huge part of the SSE (or just about any transformer output amplifier) and one that is less likely to be changed out at a later time. The one electron units are very nice, and if they fit your design requirements, I'd hold out for them, if it were me. As it is, I waited for the CXSE iron when I was originally planning on using the GXSE after being persuaded by the many reports by George and this forum.

Alternately, Edcor does make a 5k - 6 GXSE to split the difference. But I would try to get a hold of the impedance plots for each speaker if possible. My 8 ohm nominal fostex drivers are above 8 for the entire curve (although flat along for much of it). If either the 4 or the 8 ohm speakers are above or below, it will maybe help answer your dilemma.

[tbobbyt]

Ok, thanks for the info everyone. Yah that’s right the UBT-2’s don’t have the UL tap on them. Well good, I couldn’t afford them right now any ways.

Tomchr, on your B+ calculations you show a 375-0-375 with tube rectification ends up at about 400V B+? Wow, so a 350-0-350 would end up with about 370V B+. That’s about the middle to lower end of the Computer Simulation cart on the Tubelab site, interesting.

You know that 6 ohm OPT idea sounds like something maybe I should look into. I see elsewhere on diyaudio someone has put together a Simple SE kit and that’s what they use. A 6 ohm OPT. And as spopepro points out, Edcor just happens to have one in their GXSE lineup.

As for as trying to get an impedance plot for my set of speakers, I can’t find anything online. And I’m not sure where else to look. The KLH 20’s are from the 60’s. Then again I’m not even sure how efficient they are. I think I’ve seen somewhere that the Simple SE needs fairly efficient speakers. Is that true? Would an old set of KLH 20’s be worth running on the SE. I did read that the KLH 17, which is the 8 ohm brother of the 4 ohm 17’s was initially designed for tube amps. However there are a lot of different types of tube amps.

Thanks again everyone,
Bob

[boywonder]

Back in the late 60's my brother bought a KLH Model 20 system. First real hi-fi I've ever heard, and I thought it sounded amazing. I was about 8 years old and totally impressed. I imagine the speakers aren't the most efficient since they are acoustic suspension/sealed.

A few thoughts.....You could pick up a pair of KT88's to use with the 4 ohm speakers on the 8 ohm tap, they should be reasonably happy with a 2.5K load, although a more cost effective alternative might be to pick up a nice pair of cheap 90db+ speakers on your local Craigslist, like some vintage Polks, Klipsch, etc.

The KT88's would also be happy with a higher B+ voltage, and provide a couple more watts output. As mentioned, you can always add resistance to the PS to bring down the B+ if needed.

[BillEpstein]

One pair of my speakers have a 4ohm tweeter; I'm planning to build another pair with the same tweeter but a 4ohm version of the woofer. My simple SE is running 6550Cs with Transcendar 5K OPTs.

Duncan shows the following data:

Class A Va Vg2 Ia Ig2 Zout Pout

400 225 87.0 4.0 3000

I don't understand what the 2.5K output with the 4 ohm speakers mean to me in relation to this data.

[rknize]

I'm not sure what you are asking. Does your OPT have a 4-ohm tap? Will the speaker have a crossover? If you put a 4-ohm load on the 8 ohm tap, that will reflect a 2.5k load to the output tubes. This will tend to increase the amplifiers maximum power output and the distortion levels. Take a look at these tables:

Tubes & Transformers

Scroll down to the KT88 section. Look at the 5th and 6th groups of values. Those are for 5k and 3k loads at a B+ of 400V. Choose any row in each of those two groups and compare them. For example, the first row in each group:

Type	 B+ V 	Load Z	 Vp	 Vk	 Rk	 Ik	 Diss	 Pwr	 2nd	 3rd	 DF
KT88	 400	 5000	 375	 28	 220	 126	 44	 3.03	 3.3	 0.3	 4.92
KT88	 400	 3000	 388	 30	 240	 123	 45	 4.83	 4.5	 0.3	 3.55

See how the power output ("Pwr") and 2nd and 3rd harmonics all go up when the load impedance goes down?

[BillEpstein]

Originally Posted by rknize

I'm not sure what you are asking. Does your OPT have a 4-ohm tap? Will the speaker have a crossover?

Scroll down to the KT88 section. Look at the 5th and 6th groups of values. Those are for 5k and 3k loads at a B+ of 400V. Choose any row in each of those two groups and compare them. For example, the first row in each group:

Type	 B+ V 	Load Z	 Vp	 Vk	 Rk	 Ik	 Diss	 Pwr	 2nd	 3rd	 DF
KT88	 400	 5000	 375	 28	 220	 126	 44	 3.03	 3.3	 0.3	 4.92
KT88	 400	 3000	 388	 30	 240	 123	 45	 4.83	 4.5	 0.3	 3.55

See how the power output ("Pwr") and 2nd and 3rd harmonics all go up when the load impedance goes down?

First, thanks, as always.

My OPT only has and 8ohm tap, and the speaker, a first order crossover, just a cap on the tweeter.

I do see the increase in distortion and power, just the reverse of when I ran 8 ohm speakers off the 4 ohm tap of an amp that I knew reduced the power and lowered distortion. (also increases the damping factor, I'm told)

What I'm trying to understand is how the tube parameters may be used to optimise the relationship between tube and transformer, or at least understand what they signal to the builder.

[rknize]

Decreasing the plate voltage while increasing the plate current will reduce the optimum load impedance (while also decreasing output power). I'm looking at the GE 6550-A datasheet right now. For a "Class A1 Audio Amplifier", they give two examples. One has a plate voltage of 250V and a plate current of 140mA. Into a 1.5k load, this gives 12.5W at 7% THD. The other is 400V @ 87mA into a 3k load for 20W @ 13.5%.

is that what you were asking? These are just examples of what is possible. Morgan Jones' book explains how to use the plate curves to determine the optimum load impedance for any operating point of a given tube.