Lundahl LL1689 help!

[nikongod]

Hello Head-Case

I am working on an amp using this transformer in a PP output from a 6h30 tube.

I would like to know if anyone has had experience with wiring the secondaries in various series/parallel configurations as indicated on the datasheet for different stepdown ratios to drive different load impedances

The schematics & articles I could find which use this transformer for headphone use (the Raven, and a few from Andrea Ciuffoli) dont make mention of using the secondaries this way, where other a few others with sowter & "unknown to me iron" do.

I think its something worth exploring, but would like to know ahead of time if its a waste of time.

[dsavitsk]

Why would it be a waste of time? Kevin Carter does this all the time or headphone use.

[nikongod]

Why would it be a waste of time? Kevin Carter does this all the time or headphone use.

I found a thread about the Raven in the K&K section of audioasylum

here

that commented on a some people being unable to get enough gain out of the thing without the input/gain stage which got me second guessing the feasibility of running the secondaries like I want to (Reading it a bit more the OP may be based on the LL1680 transformer which explains a lot of that). I guess different peoples' standards of enough gain may differ. Gain of 1/2, 1, and 2 is plenty for me for 32, 150-300 and 600 ohm headphones respectively.

After seeing that I was curious if this may not work out well for some reason I had not considered. I can only see it going well but maybe I have tunnel vision from my enthusiasm. I like how switching secondaries works as an effective "gain switch" matched to the target load, I dont need that much gain that I need a gain stage, and the increasing output impedance from setting the transformer up for higher impedance loads dosnt bother me.

Did I ignore some key point? Wiring the switches to switch the secondaries will be an interesting challenge (I think it will be the crux problem of this amp), but does anyone think I missed something that should be consided?

[dsavitsk]

Did I ignore some key point?

Two things to think about. With all the secondaries in parallel, the copper impedance is a little more than 4 ohms. No big deal. When you step up to the next level, the copper impedance is about 16 ohms. You will, in theory, be jumping up 6db, but if you run it into 32 ohm phones, ~3db of that will be lost due to the copper so the gain will not be as big as you think. This is less of an issue with high Z phones, but it exists there, too.

Second, keep in mind that the tubes only see 1/4 of the primary impedance in a push pull amp. 18:1 is ~10K:32 which is like loading the tube with 2K5. Since rp is ~1K5, this means that Zout is likely higher than you might think -- especially when you all in the 4 ohms from the copper. This is not necessarily a bad thing -- as I've said elsewhere, I think it can actually be a useful thing, but it is worth being aware of as, if nothing else, you'll lose some more gain there, too.

The TCJ PushPull Calculator might prove helpful. It isn't spectacular, but it is definitely worth the $30 as it will help you get things right faster than trial and error.

If I were doing it, I think I'd use a quad of 6C45's instead for the extra gain and lower rp. Or, I'd use some step up's (LL1674) on the input.

[nikongod]

I have been reading more about this, and from what I have found the impedance seen by the tube depends whether its in class-A or class-B. class-A gets 1/2 of the whole primary, class-B gets 1/4.

I thought some more about the gain & amp in general, and Im not sure where to go there. My plan at the moment is to use a CCS to feed the center tap of the transformer (with no shunt reg, I dont think it will be necessary here), with each triode set up as a common cathode with its own resistor and cap. This will buy me a little gain compared to my original boring plan of a differential pair with the CCS tying the 2 cathodes together.

Im not sure I like cathode LED's but I'l probably try them too. a headphone sized red-light-district amp could be fun.

I thought about the 6c45 but I am frankly afraid of them.

I got TVC's with +6, +12db taps from an estate sale. I hope I have gain covered.

[aerius]

Gain shouldn't be an issue with Grados, I'm running a set of 6BL7's in P-P with a Lundahl 1663 which is a 5k:8 ohm transformer to drive my RS-1. On the other hand my CD player has a variable output which goes up to 5V and most of my listening is in the 60-70dB range. My amp is LED cathode biased, it was CCS'd via the transformer centretap but I stole the IXYS 10M45S's for another project, I can't say I notice any difference since my PSU is ridiculously overbuilt.

The 6C45 is definitely a step up in difficulty but as long as you have a 'scope you'll be fine. You just need to pay real good attention to your wire routing and use grid stoppers to keep the damn thing from oscillating. And don't forget to tie all the cathode & grid pins together (there are 4 & 2 of them, respectively), leaving one or more floating invites problems.

[nikongod]

I spent some more time looking at the 6C45pi tube curves. It is very attractive for its absurdly absurd absurdity. I really like how linear it looks into a 6.6Kohm load (40 ohm load on the 18ct:1 setting) at 150v/30ma idle. I am a little put off by its very high gain and input capacitances because of miller effect though. I think this tube would work really well if a 10K:600ohm input transformer were used with a low impedance attenuator between it and the tubes, but Im petty set on the attenuator I have coming.

For a later date How do people hook up the cathodes of the 6c45 in a longish tail pair? The bias point above only needs Vgk of ~1.4V which dosnt buy much in the way of a shared cathode resistor after the stopper resistors are considered. I considered a CCS to a negative supply, but that prevents the use of a CCS to feed the CT of the transformer. Maybe a big fat plate choke for up there. Moar iron=moar better? Maybe its just superfluous in a forced balanced amp. I'd certainly appreciate more info on this.

I found the circuit I was trying to describe BTW. Im going to actually bread-board this one like a good boy, and I guess I'l decide then what Im going to do with the cathodes.

Im sticking with the 6h30 for this one, but I do appreciate the help and advice.

[aerius]

For a later date How do people hook up the cathodes of the 6c45 in a longish tail pair? The bias point above only needs Vgk of ~1.4V which dosnt buy much in the way of a shared cathode resistor after the stopper resistors are considered. I considered a CCS to a negative supply, but that prevents the use of a CCS to feed the CT of the transformer. Maybe a big fat plate choke for up there. Moar iron=moar better? Maybe its just superfluous in a forced balanced amp. I'd certainly appreciate more info on this.

LED bias, but you'll need to put 2 or 3 LEDs in parallel to get enough current capacity and a small value resistor in front of each one to ensure even current sharing on the LEDs. Same as the picture you linked, just replace the cathode bias resistor with LED bias.

There's a couple DIY amps I know of which use this arrangement with a 2:1 (or was it a 4:1?) input transformer in front of the tubes, and a TVC in front of that to set the volume. Using a pot as the volume control will likely require an additional cathode follower stage to drive the miller C on the tubes.

I've only played with the 6C45 in SE parafeed mode, it worked but it was darn finicky and I don't have a 'scope so I didn't explore it much. The tubes went back into the parts box and I went back to easier stuff.

[dsavitsk]

LED bias

I think I disagree -- the CCS load is good as it is a high impedance load that keeps the signal current away from ground. An LED is a low impedance load that shunts AC to ground, exactly the opposite of what you want. I'd just stick with a resistor. A negative supply to tie it to would be helpful. If you are feeling really ambitious, use the choke in the cathode. Otherwise, use a CCS in the cathode and a shunting element as a voltage regulator (with a cap from B+ to the cathodes).

And, I am not sure I'd run a 6c45 as high as 30mA. 20mA has worked better for me.

[aerius]

I think I disagree --

Yeah, you're right since he did mention he wanted a longish tail pair. Problem is a good LTP can't really be done with the 6C45P using resistors unless you use a negative supply, and if you're going to use a negative supply then you might as well CCS it. I'm not sure a choke on the cathodes would be a good idea with these tubes since they're prone to noise pickup & oscillations, it seems like asking for trouble to me.

Personally I don't think a LTP would be the first choice with the 6C45P, and it definitely wouldn't be my first build using the tube. I'd lean towards using a transformer phase-splitter to feed the tubes and run them as a simple P-P pair, I give up a bit in signal balance between the tubes but it makes the rest of the design & build a lot simpler. LED bias on the cathodes, CCS feeding the centre tap on the transformer, and a cap from the centre tap to the cathodes.

And, I am not sure I'd run a 6c45 as high as 30mA. 20mA has worked better for me.

That's what I've heard as well from various sources, I ran mine at a bit over 20mA and it seemed to be pretty happy there. Apparently it gets pretty temperamental at higher currents.

[nikongod]

Hello Head-case

I (still) have not built the amp I'm interested in here, but have a tangentially related question that I think may be of interest to several other people who have expressed a desire to use this transformer in their own amp projects.

I am interested in making a fairly compact PCB to switch between the various secondaries on the transformer.

To understand this you will need to download the LL1689 datasheet from K&K audio.

Here is the general circuit.

The general operation of the circuit is as follows:

The top 2 switches put windings 3&4 in series or parallel with 2&5 and likewise for 10&7 and 8&9

The third relay puts winding 10&7 in series/parallel with 3&4

I am pretty sure that with these 3 switches I can get all 3 possible outputs as indicated in the datasheet.

The bottom switch is somewhat optional. It is there ONLY to switch the headphone or preamp outputs. Maybe not appealing to everyone, but its only a nominal additional expense.

I plan to control this mess with a 4-pole-4 deck switch. The switch will only be used to switch DC (at low voltages & currents) so it should not be hard to find a truly cheap unit I did consider using some sort of SS logic chip, but gave up.

Anyways, I would appreciate any corrections to my circuit above.

Chapter 2:

Im pretty sure my circuit above is right so I drew the circuit in the EAGLE schematic editor. Here it is:

I think I copied the schematic over properly into Eagle, but greatly appreciate any corrections. NOTE! The second relay is "upside down" in this schematic compared to the hand drawn one.

I planned to run this board off of spare current in the heater supply on my amp. I think 2 channels will only use 100mA, so Im happy. I could not find 6V relays to match my heater supply but 5V parts are everywhere: the resistor is to drop a volt.

Chapter 3 (finally)

I let Eagle make me a double sided PCB. I wasnt sure how to get EAGLE to make a single sided PCB. With a bunch of trace re-routing I got the while mess on one side of the board.

This whole thing now is dependant on both schematics being right... But even if not do you have any recommendations for this PCB?

Im not sure how large it is (heh) or how to put holes in the corners. For size I would ideally like to have the whole thing only slightly larger than the transformers so they can sit next to or underneath them in the amp and not waste a lot of room.

Thanks in advance for any help you can give

Edited March 10, 2011 by nikongod

[luvdunhill]

Perhaps adding diodes to limit the back EMF off the relays (assuming you're not worried about the time it takes to latch and unlatch)? Then again, I assume these aren't going to be firing all that often, huh.

Edited March 10, 2011 by luvdunhill

[nikongod]

Thanks for the advice to add the diodes. I plan to switch these with a mechanical switch, the diodes seem more important to SS controlled gear, but I'l add them anyways.

I dont think that the switching time will be very important to this. The circuit is just being used to switch the secondaries of the transformer to accommodate different headphones.

I found lots of information on protection diodes, and many mentions that the diodes slow things down, but could not find any way to estimate how much. Got a link? It sounds like the diode delays the unlatching of the relay until it bleeds off the voltage spike, usually a few miliseconds. Im totally OK with that.

Any thoughts on the basic circuit?

[digger945]

My bad Ari. I didn't click on the link to the other K&K PP amp in your second post. The CCS on the outputs was what I was talking about.

Is that amp in your second post (post #3), is that what the Raven is?

Edited March 12, 2011 by digger945

[Pars]

Chapter 3 (finally)

I let Eagle make me a double sided PCB. I wasnt sure how to get EAGLE to make a single sided PCB. With a bunch of trace re-routing I got the while mess on one side of the board.

This whole thing now is dependant on both schematics being right... But even if not do you have any recommendations for this PCB?

Im not sure how large it is (heh) or how to put holes in the corners. For size I would ideally like to have the whole thing only slightly larger than the transformers so they can sit next to or underneath them in the amp and not waste a lot of room.

Thanks in advance for any help you can give

Hello Ari! (to paraphrase your opening)

Holes: just use the Hole command (lower left in the toolbar) and size it appropriately. I'm using 4.445mm right now on a board and it about right for a 4-40 screw.

As for the board dimension, use the info (i) command on each of the lines. In the dialog box, you can resize these as you see fit. I always make sure the lines are straight as well (yours aren't) by making that dimension identical.

Schematic looks right from what I can tell, but I agree about the diodes. E=You could just use 1N4148s or the like for a small one. I'd never let Eagle do the routing, but this looks OK.