K&K Audio / Lundahl Transformers

Welcome! Need support, you got it. Or share you ideas and experiences.

Return to K&K Audio / Lundahl Transformers


Best (least bad?) misuse of LL1682

98.206.37.100

Posted on October 15, 2009 at 12:26:56
Tom Ronan
Audiophile

Posts: 13
Location: Chicago, IL
Joined: October 24, 2005
Hi All,

A couple of years back I bought a pair of LL1682 transformers for use in a low-powered low-current 845 amplifier driving Altec 755As. It was an odd project, but came out beautifully. I ran the metal plate 845s at 450V and 50mA with a bias around -45 if I recall, driven by EF37 pentodes similar to a WE91A circuit. The Altec speakers were 4 ohms, so the normally 5.5k:5ohm LL1682s presented ~4.4K load to the 845s. The sound was surprising - the bass was really strong and powerful (kind of the opposite of what I expected). If the amp had a flaw, it was a little less delicate than I had hoped for, but that could have been due to the whimpy pentode driver tube, but all-in-all it was a very nice sounding amp.

Since then everything has changed, and I now have an 8 ohm system. I would like to put these transformers to use again. I have two potential uses in mind, but due to the 8ohm requirement, both uses 'tweak' the ideal situation for the transformer and cause it to be used in less-than ideal ways. Which do you think will have the least impact on the sonics?

Option 1: I really enjoyed the sound of low-powered 845, though I would probably do without the low-current pentode input stage this time. Now, with 8 ohm speakers, choosing the load is more complicated. The LL1682 is a 16+16:1+1 transformer. With secondaries paralleled, it presents a ratio of 32:1. With an 8 ohm secondary load, there would be an ~8.2K primary load presented (8 ohms * 32^2 = 8192 ohms). The amp might have put out around 4 watts before, so doubling the load to 8Kish would reduce the power significantly. I know that people 'misload' transformers all of the time, but as I understand it (from the math presented by Mike L. a long time ago on the SP mailing list) if there is a direction to err, ratio down. (In other words, make a 10K transformer 5K but putting a 16ohm speaker on an 8ohm tap, but not the other way around.) But maybe this is the least problematic use?

Option 2: The high load presented Option 1 gave me an idea for an alternate use. I have some 10/VT-25 tubes, and have wanted to build a nice amp with them for ages. Although the textbooks indicate that a SE 10 amp should have a 10K load, it's really not ideal. At that operating point, they have a 5K Rp, and a load of 2xRp is a bit light. You'd really rather see a 15K load (3*Rp). If 10k SE transformers are hard to make with decent bandwith, a 15K transformer would be really tricky. However, a paralleled 10 amp would only need a 7.5K load with about 40mA of current. So, using the above 8.2K load connection on the LL1682 into an 8ohm load, I could build a paralleled 10 amp with maybe 2 watts output. Still light on power, but it would be a 10 amp which would be cool.

Option 3: Not entirely ready to give up on the 845 plan, I did some calculations and came up with a 4W 845 option, but maybe at too much of a performance cost. If you wire the LL1682 secondaries in series you have a 16:1 transformer. When loaded with 8 ohms would present a 2k load to the 845, which is too low. However, if you call the secondary 16 ohms, then the CT of the seriesed winding is an 8ohm tap. Using a 16 ohm speaker on the whole winding, or an 8 ohm speaker on the center tap would reflect a 4K primary impedance to the 845. But I don't really know how bad a hit will the HF response take by connecting the secondaries in series. Does anyone have an idea? Of course, if I choose this option and it sounds like poop, I can always wire it as in option 1, but I would appreciate any thoughts or math you wish to weigh in with to help me predict in advance.

Thanks for your thoughts and time!
Tom

RE: Best (least bad?) misuse of LL1682, posted on October 17, 2009 at 09:57:55
Tom Ronan
Audiophile

Posts: 13
Location: Chicago, IL
Joined: October 24, 2005
My one reply has disappeared. (thanks for the encouragement though!)

Maybe I can rephrase my question more simply:

Which is worse for frequency response/sound on LL1682 transformers?

(a) 8 ohm speakers on the 5.5 ohm tap, or
(b) putting secondaries in series and using only 1/2 the secondary winding

Or are my concerns academic and it will sound find either way?

If no one has any input, I'll just build the 845 amp so I can swap both ways and I'll report back.

RE: Best (least bad?) misuse of LL1682, posted on October 17, 2009 at 10:19:59
KevinC
Manufacturer

Posts: 1999
Joined: April 19, 2001
Contributor
  Since:
October 7, 2002
I would guess that 8 ohm speakers on the 5.5 ohm secondary would be preferred.

Kevin Carter
K&K Audio
www.kandkaudio.com

RE: Best (least bad?) misuse of LL1682, posted on October 17, 2009 at 22:46:49
Tom Ronan
Audiophile

Posts: 13
Location: Chicago, IL
Joined: October 24, 2005
As I said in my first post, I was a bit concerned about the power loss when putting such a high load on the tube (but gaining lower distortion in the trade-off). I ran some numbers tonight and here's what I came up with:

Power Out 3.16 W
into impedance 8 ohms

Voltage needed (at speaker) 5.03 V RMS
OPT pri 8200 ohms
Ratio 32.0

Voltage Needed (at plate of 845): 160.97 V RMS

845 mu 5.5
845 Rp 1700
845 Gain 4.56

Voltage Needed (at grid of 845): 35.34 V RMS = 99.96 V p-p

That means that with an op point of -50V bias, and an ~8k load, I can still get over 3 watts out. (FYI, the same calculations run on a 4100 ohm load yields 4.6 watts with the same -50V bias.) I was expecting more power loss, so that's a nice surprise.

Three watts is plenty for me, and I like the idea of the lower distortion. Looks like I'll go the 8K route, but I may still give a quick listen to the seriesed/4K configuration just to hear the difference.sonically.

Thanks, Tom

Page processed in 0.040 seconds.