Bottlehead Forum
Bottlehead Kits => Eros Phono => Topic started by: Davenportt on July 07, 2011, 08:41:06 AM
-
Hello all,
Back in February there was a post about how the self-bias circuit works and how to remove it. Can anyone tell me: 1.) If one uses a high enough quality NOS EF86, what kind of sonic improvement are we looking at by eliminating the self-bias circuit? and 2.) How do I know if my NOS tube is good enough? That is, is it safe to find out whether it biases up correctly by merely plugging in different tubes and measuring voltages? Do I run any risks to other components if I just plug in a NOS EF86 and it's NOT up to snuff?
Thanks in advance,
David
-
If you use a tube that won't bias properly like an EH it will not harm anything as a temporary trial, but the plate voltage will be off and the 6DJ8 will not be biased properly either. Sonically it is just a tiny, tiny bit smoother on top to my ear when you have a tube that allows use of the LED instead of the servo circuit.
-
Dan,
How old does the NOS tube need to be? Audio Cheapskate that I am I have a dozen of the Russian tubes from the 60s and 70s. Very good decades for me, hopefully you too.
-
The "real" EF86s/6267/EF806S we have found to bias up nicely with the LED - Mullard, Amperex, Telefunken, Valvo, etc. The older RFTs and the Soviet winged C 6J32Ps seem to work fine also. The EF86s that don't seem to bias up properly with LEDs are the new production EHs we use in the kits. But we have seen the oddball sample here and there of other brands that didn't work, so you just have to smoke it and see.
-
Is there a voltage measurement that tells that you have the right bias, say at the cathode?
-
Not at the EF86. The cathode voltage is fixed by the LED at 1.56VDC. The problem is the plate voltage can go too high with some tubes when biased at 1.56 V and that will screw up the bias of the direct coupled 6DJ8, so the EF86 plate voltage is what you should be measuring.
-
Looking at the manual OB connects to the EF86 plate and should be 100V.
I'm looking through the threads today and thought I might post the voltage.
-
To clarify, you cannot make a useful measurement on a working stock Eros, because the servo bias will correct the plate voltage to 100v. You have to convert to the LED and then measure the plate voltage, and try another tube if it is off. I think you want between 95 and 105v at the cathode - it's pretty tight.
The really important measurement is the difference between the 6922 cathode and plate. It should be between 60v and 80v, so that this triode has adequate bias and headroom.
There is a fair amount of DC feedback from the 6922 cathode to the EF86 screen grid, which was intended to keep the parameters close enough. Works for the classic EF86, at least most of them, but it's not enough for the modern variants. I think we did try a few others along with the EH.
Basically, if you can't afford to throw away an expensive NOS EF86 now and then, you should not make this modification. Stressing about the tubes and their cost detracts from the listening experience!
Incidentally, I don't know why there is this difference. I suspect it may be differences in the metallurgy (contact potential) of the grid and cathode materials, since the dimensions of the parts are set by the machinery, and the modern reproductions pretty often use the same machines.
Incidentally, I have found big differences in 6922/6DJ8/7308/6C23Pi grid current vs. voltage, which affects the minimum bias needed on that tube. The 6DJ8s have the greatest problem, so I suggest you avoid that particular variant - the Eros is designed around the 6922, and the Repro high end preamp assumes you have found some good 7308s.
-
PJ,
Thanks for the very complete explanation. I was reading over Eros threads yesterday and it dawned on me that the target voltage hadn't been mentioned.
I should have added that the target was for a converted Eros. I don't foresee that I would make this conversion myself. I'm not sure the lower noise level is something I could hear anyway with the ringing in my ears.
-
A couple of weeks ago, I removed the self bias circuit by replacing the 2700uf 4v cap with a HLMP-6000 diode, replacing R3(176 ohm resistor) with a jumper wire and removing the 2n2222A transistor. The Eros sounded fabulous before the mod. Fortunately, this same great sound did not change after the mod. I also replaced the 100uf 160v electrolytic 6922 cathode bias caps with 47uf 160v obbligatto's. Before plugging the Eros in, I used E.A.R. viscodampers from Rathbun to mount the nearly full range drivers. This combination made an improvement on the soundstage-larger, more focused with slightly less grain. Paul
-
I made this same change some time ago, but I wasn't sure if talking about changes to the circuits was ok on the forum. I'm glad to discover that modifications like this one can be shared here.
The change gave me a big scare about a week after I made it, or so I thought. All voltages had settled in just fine with my tubes at first, but I decided that checking was better than trusting and put my trusty Fluke DMM to work again a week later: oh no, 145V or so on the cathode, more than 200V at OA ... change tubes ... even worse ... something like 160V and 240V ... oh no, undo the mod, transistor back in, cap back in, resistor bridge cut ... what? ... the plate voltage still isn't back at about 100V ... and then the display on my voltmeter went blank, because the battery was completely dead by now.
I replaced the DMM battery and everything was just peachy, except now I had gone back to the original circuit design. So then I put the modification back in and checked everything on more time. After all that commotion I was extremely pleased to find everything working perfectly: 96.8V / 170.5V in one channel and 97.1V / 165.8V in the other. The cathode to plate difference in the 6922 is well within range at 73.7V and 68.7V respectively. Whew!
It ended up looking messier than I like it and everything took much longer than it takes to read this message, but maybe I learned something.
A couple of weeks ago, I removed the self bias circuit by replacing the 2700uf 4v cap with a HLMP-6000 diode, replacing R3(176 ohm resistor) with a jumper wire and removing the 2n2222A transistor. The Eros sounded fabulous before the mod. Fortunately, this same great sound did not change after the mod. I also replaced the 100uf 160v electrolytic 6922 cathode bias caps with 47uf 160v obbligatto's. Before plugging the Eros in, I used E.A.R. viscodampers from Rathbun to mount the nearly full range drivers. This combination made an improvement on the soundstage-larger, more focused with slightly less grain. Paul
-
The 160V capacitors in my Eros are still the originals, but I will probably try that modification also.
The Eros is plenty nice no matter what. Thank you for a very well thought out design that works beautifully right out of the box ... some assembly required.
-
Every so often I get odd readings and it tells me that this year I need to change the battery in the Fluke. I trust the readings I get from it and when it makes no sense, it is the battery.
-
Switching to LED bias for the EF86 tubes in the Eros is the best change I have ever made to a Bottlehead circuit. I have listened to 3, maybe even 4, times as many LP's as I did before the change for the last three months. Right now it's Ry Cooder's Bop Till You Drop. I had forgotten about that one ... thrilling.
Getting the channel balance back to minimal differences took a bit of doing. Everything messes with everything in phono preamps: bias, channel output, RIAA curves, ... they are heaven for the OCD challenged.
Do this, if you have the tubes for it and if you have the patience to get it right. The result is shockingly good! The music will jump out at you.
-
Hello All,
I recently switched to the LED bias, to hear what all the fuss is about. I have several NOS types, Teles, Philips etc.
Most of the voltages are spot on with the only concern being the 6922 plate to cathode voltage being only 48 - 52 volts, depending what tubes are in. IIRC the 170V has always been a shade low - around 150 volts - from the original build. I read earlier in the post where PJ mentioned the plate to cathode should be around 70 volts.
I guess my question is: Is it worth being concerned over and are there any recommendations? Would changing the value of the 47K resistor help any?
I hear what the fuss is about, by the way.
Thanks in advance
-
Just checking here, the diode you used, is it this one:
HLMP-6000
http://ca.mouser.com/ProductDetail/Avago-Technologies/HLMP-6000/?qs=sGAEpiMZZMs4quMj8r4lmvKuq%2flfwXl0ro7P6wPamrM%3d
I could use a few.
Thanks,
-
The Avago HLMP-6000 types [that's the tiny Bottlehead LED] has been giving me the lowest bias voltage values. For most tube triplets [2xEF86 and 1x6922] that has resulted in the best combination operating points, i.e. something like 95V plus for the EF86's and about 160V for the 6922.
It still took some swapping around to find combinations that bias up properly.
The setup that I have right now actually uses standard red LED's, because the lower voltage of the HLMP-6000 [about 1.56V] had the EF86 tubes at 107V and the 6922 at about 150V. The standard red LED gives me about 1.65V.
Try measuring your bias voltage with the servo still in place. That should actually give a pretty good indication of where you need to be, if I'm not completely mistaken.
-
Thanks, I will do that. I have been meaning to buy a hand full of these so I will go ahead. I have a number of old red LEDs I might give them a try also.
And thanks for the voltage targets. You saved me a lot of digging.
-
I think a light may have finally gone on inside this dull head of mine.
Thank you
-
Yes, it's worth being concerned. No, don't change the resistors - it's direct coupled, and every change you can think of affects both tubes - you have to achieve the desired voltage and current in both tubes all at the same time in order to get proper operation - that is, correct equalization, correct gain, and correct (low) distortion.
If the 6922 plate to cathode voltage is low, then the grid to cathode bias voltage difference is also too small, and grid leakage current provides a shunt resistance which reduces the maximum gain of the first stage. Bottom line - improper equalization and less bass. Each individual 6922 or whatever else you have in that position will have a different threshold of critical leakage, and it will change with age - the design operating point assures that most tubes will work correctly most of the time. This is exactly the reason we went to a servo bias - so that most tubes would not exhibit this problem.
-
Thank you for clarifying this further for me. I wish I knew more about the theory behind it all, maybe I wouldn't get hung up on every mod I read about! You pretty much nailed it with the equalization - while the highs are more open and detailed, the bass is a tad thinner. I'll likely go back to the servo config as the amp is amazing as designed - and had been trouble free with pretty much any tubes I plugged into it.
Thanks again
-
Hmmm, well mine isn't working so well. The 6BR7's are only getting around 70 to 75v and the 6922 plate to cathodes are 114v and 112v. Is it me or the 6BR7's? Am I going to have to convert these over to EF86?
-
You might be able to get there by changing the type of LED that you are using. Different types of LED have different forward voltage drops. What kind of LED are you using now?
-
I'm using the standard Avago HLMP-6000 that Bottlehead provides with the kits. Could you specify a part that might work better and a voltage range I should be looking at? LED specs get a little confusing for me. Thank you!
-
I use small indicator LED's with low light output (mcd) values just so they don't light up the whole room from underneath the chassis.
You should measure the voltage across the HLMP-6000 so that you know where you are at now. Other red LED's will give you a very similar voltage. Green ones are usually the next step up in forward voltage, then yellow. I haven't looked at specs for blue or white LED's in some time. Another interesting type of LED are the IR (infrared) ones. They typically have a forward voltage drop around 1.2V.
So basically what you can get is: just around 1.6V with the HLMP-6000, a bit more with other red LED's, 1.7+V with small green indicator LED's, 1.9V with yellow ones, 2.3+V with two IR LED's in series. There must be other LED's out there to fill the gap between 1.9V and 2.4V I just haven't had to look.
How easy is it for you to go back to the servo circuit on one channel? It would be helpful to know what your bias voltage is with the servo circuit in place.
-
The HLMP-6000 was selected some years agon by John "Buddha" Camille, for its exceptionally low noise and its low dynamic impedance. Other diodes will have different voltages and impedance - by the way, there are several different red formulations.
The direct couple circuit has some negative DC feedback to stabilize the operating point, derived by feeding the EF86 screen from the 6922 cathode. (Incidentally, the 'Tode guitar amp under development has the same circuit - an old Mullard innovation.) Most old-stock EF86s are close enough to the spec that they will bias up correctly with the design choice HLMP-6000. Modern production EF86s are less consistent and many need a greater amount of DC negative feedback, provided by the servo circuit.
Obviously tubes that differ further from the EF86 specs can sometimes be made to work by the servo, but would be especially unlikely to work without it.
An alternative to the servo bias is adjustable bias, in the present case substituting other diodes until the right operating point is achieved. Unless you have access to tube curves at 100v on the screen, it is impractical to predict what the bias should be - I had to develop my own EF86 curves to do the original design. I have another adjustable bias in mind, but have not had a chance to test it yet - and I won't share it until I know whether it works, sounds good, and is quiet. Risks of changing the bias scheme are noise and distortion. Risks of changing tubes are noise and RIAA equalization accuracy - the tube's plate resistance is incorporated into the equalization, primarily affecting the 50Hz low frequency corner.
We elected not to pursue such a design because it requires re-adjustment as the tube ages and whenever tubes are replaced, including rolling tubes. It's not just the practical difficulties, I also worry that too many Eros' will be operating out of adjustment, sound bad, and create a poor opinion of the Eros.
-
Thanks guys. I will rewire one of the channels for EF86 and restest. With the Tube Phono at home I've collected a few that should work nicely.
-
Alright, a little update. As I've got a stash of these 6BR7's I reinstalled the servo's and measured. Not good... Here is what I'm getting with the 6BR7 and 6922: 52.7v/96.7v, 58.3v/ 94.2v. As it stands, none of my voltages are w/in range. So I converted one channel to EF86 and restested; line voltage is 119.6v (50Hz)
With servo: 60.2v/ 95.4v
With LED and Winged C EF86: 84.6v/85.1v
With LED and CV4086: 61.3v/95.1v (this voltage changed for some reason after I swapped tubes and never returned to this level)
Edited to save on posting. I converted the other channel over and have found that not all tubes are the same. Here are the final results with CV4086 tubes (the best of the four):
R 71v/ 89.3v on EF86 cathode
L 70v/ 93.3v on EF86 cathode
PJ, my voltages are good for the 6922 but a little low for the EF86. Do I need to keep trying tubes until I get it right??
-
Argh...darn me for reading this forum and getting obsessed about mods. I just HAD to try it. Ugh.
So I have a pair of Telefunken EF86's and a spare pair of the Avago LEDs, so I figured I would give this a try. The actual mod was easy and went very smoothly.
My plate voltage for the EF86 (measured from OB to ground) is now ~105v (vs. ~95v with servo). My 6922 plate is now ~150v and the 6922 cathode is ~105v (vs. ~160v and ~97v with servo).
So, my EF86 plate voltage is in-range, but the 6922 plate to cathode gap is now too small (right?). This may be a noob question, but does this mean my LED bias voltage (1.54v) is too high or too low, e.g. do I want something closer to 1.3 or to 1.7v?
Profuse thanks in advance. BTW, the Eros ROCKED until I started messing with it. Thanks for an awesome kit!
Drew
-
You'll need a more specific voltage reading of the cathode voltage of the 6922 and the plate voltage of the EF86, see if you can grab the first decimal place and report back (you can also listen, it's very audible when things aren't in check).
-
You'll need a more specific voltage reading of the cathode voltage of the 6922 and the plate voltage of the EF86, see if you can grab the first decimal place and report back (you can also listen, it's very audible when things aren't in check).
Thanks, CB. Here are the more precise measurements:
EF86 Plate (at OB) = 103.5v (Socket A) / 101.5v (Socket B)
6922 Plate (at OA) = 149.2v (A) / 149.5v (B)
6922 Cathode (at breg) = 104.3v (A) / 102.4v (B)
BTW, it sounds fine, but PJ's post on the 6922 voltage gap made me nervous.
Thanks again - Drew
-
A little higher than 1.54V will get you closer to the optimal voltages, if I remember correctly. Not much higher though, it is a trial and error thing, but worth the effort.
-
Those numbers look acceptable. What I often see is that the EF86 plate is slightly positive with respect to the 6922 cathode, which is seriously undesireable, or that the gap is overly tight (a few tenths of a volt or so).
-
Grufti wrote:
The setup that I have right now actually uses standard red LED's, because the lower voltage of the HLMP-6000 [about 1.56V] had the EF86 tubes at 107V and the 6922 at about 150V. The standard red LED gives me about 1.65V.
[...]
Grufti,
Can you provide any further detail about the "standard red" LEDs that you successfully used (brand, model/part number, etc.)? I think I'm in a very similar boat where the 1.54v voltage drop of the HLMP-6000 is just slightly too low for optimal biasing. I went on Mouser and Digikey to look at red LEDs and was overwhelmed (and confused) by all the options. Any further guidance to narrow my search is much appreciated.
-Drew
-
Are you saying that there is a problem in that the Eros sounds bad now, like distortion? Or are you just worrying about numbers? As PB said, if the 6922 cathode is positive with respect to the EF86 plate by an adequate amount a little difference in the 6922 plate voltage doesn't matter that much. Those numbers can change a little with every 6922 you try. If it bothers you, put the servo back in. That's what is controls.
-
Are you saying that there is a problem in that the Eros sounds bad now, like distortion? Or are you just worrying about numbers? As PB said, if the 6922 cathode is positive with respect to the EF86 plate by an adequate amount a little difference in the 6922 plate voltage doesn't matter that much. Those numbers can change a little with every 6922 you try. If it bothers you, put the servo back in. That's what is controls.
Thanks, Doc. No, it doesn't sound bad...the low bass may be ever-so-slightly rolled off, but that's subtle enough that I'm probably imagining it. I'm just looking at numbers, but it isn't bothering me...just trying to have my cake and eat it, too. I'm actually enjoying the process of trying to dial this in...but I made sure I have two spare sets of parts for the servo before I started. Plan B will be to replace the servo and stop obsessing.
-
Low bass sounding rolled off is most likely just the difference in the sound of the tube itself vs. the original one. Remember that we design for a certain operating point, but that we end up measuring and reporting exactly what we measured in the prototype whether it hit the design values exactly or not. With tube circuits there is always going to be a little variation. Understanding what numbers can be different than the manual and be OK, vs. which ones are bad if they are different than the manual is really the key issue.