Cornet Problem

jameshuls · « **on:** 2 Dec 2008, 04:28 pm »

After reading the posts on Leitmos' problems below, I think it is time to make my Cornet problem a matter of public discussion! The problem? None of the tubes after the rectifier are firing up. There is some voltage on the wiring, but no heater response in any of the tubes - I have many tubes, and so have rolled them out to confirm they are not the problem - they are not! Also, there is no voltage on some of the resistors in the signal path. Now, first of all, I have tested every resistor in the unit, and replaced those that were off. I also have also tested the voltage (AC and DC) at every component. There are definitely some anomalies in the signal stage. However, before posting all this information, I would like to cut to what I suspect may be the problem and see what others make of it, as it would be an easy fix; if the transistors in Q200 were fried, could this cause the above-described problems? I appreciate any input!

More info: It seems significant that the LED goes green immediately, w/o any lag.

tubesforever · « **Reply #1 on:** 3 Dec 2008, 08:00 am »

Did you check the transformer voltages at the PCB lugs?

The voltage for the H+ circuit should be 6.5 or higher from my memory. If your Schottke Diodes are dead or inserted backwards they could effect the H+ supply line.

I would check the voltage at both ends of the R223 5 watt resistor. The side facing the RCA's on a C2 should show a higher voltage and the opposite end is the final DC voltage headed to the heaters. If the R223 resistor is cooked you would show zero volts out.

I would first verify the voltages coming off the transformer. Then move down the H+ supply line one component at a time.

jameshuls · « **Reply #2 on:** 3 Dec 2008, 03:29 pm »

Other than the voltages printed on the board, is there somewhere else where I can reference the approximate voltages one should see at each component?

jameshuls · « **Reply #3 on:** 3 Dec 2008, 03:55 pm »

The voltage at C102 and R103 is 6.3VAC. R223 measures at 18 VAC - resistance is beyond measurable with the power on. D101-4 have a measured voltage of 9.1 VAC, C106 a voltage of 18VAC. C209 and C210 have 0 VAC and 4.3 mVDC - perhaps the problem is at C106?

tubesforever · « **Reply #4 on:** 4 Dec 2008, 09:51 am »

18 volts ac at R223 stands out as a major problem. The R223 is simply a drop down resistor to drop the voltages from the diodes down to something the 12AU7 can live with. Can you check the voltage. There should only be DC voltage at that resistor. The Schottke diodes and the LC tank are there to convert the transformer's AC voltage into smooth DC for the heaters.

18 volts AC is flat out wrong.

First thing to do, make sure you plug in all the tubes. Don't run these tests without the tubes.

Second make sure your Cornet has the power selected for 120v. Check these carefully. The Black vs the Black and Red are tough to separate without good light. I also used a magnifying glass because that black/red is tough to see on mine. In fact I can only see traces of red on the white printing. You must get black and white to the line lugs on the PCB. You should have solid blue and solid brown for the neutral lug connections.

On the left side of the board you will have the secondary windings coming down to the board to feed the H+ circuit. These wires plug in as Green, Green/Yellow, and Green respectively. These feed power to the H+ circuit.

I am showing 3.00 volts on each of the three green, green/yellow, green lugs after about 10 minutes of warm up time.

The Ground is important. I like to use the phono ground lug at the RCA inputs as the central grounding point of the phono section. From that lug I run one wire to the PCB earth ground connection and one wire goes to the center IEC connection.

Make sure that you have the 10k ohm Electrolytics oriented to ground properly. Also verify the bands on your Schottke diodes are correct. If the LC tank or the diodes are backwards this might create issues. I am not sure they would produce 18 volts AC but these are the only pieces that are going to effect the dc output going to the H+ step down resistor at R223.

You should have high voltage there on the top side and a lower voltage on the lower side. What is the voltage on the lower side of R223? Also check the voltage on the center PCB leads of each tube socket. They should all be within .1 vdc of each other and equal to the lower side of R223 or I suspect you have some damaged tubes.

Back to the H+ section, I had been showing about 6.6 to 6.9 vdc at the top side of R223 with the Schottke's so the R223 is simply there to drop the voltage to something appropriate to heat the 12AU7 tubes. I like mine about 6.175 vdc or as close as I can arrange it. However when I switched to the Hexfreds this voltage dropped to 6.0 vdc. So I cannot use any resistor at R223 on my Cornet 2. I have bypassed the R223 with some high quality copper 5 nines pure copper wire. I like the Hexfreds and do not plan to remove them.

18v is too high James. If the voltage at R223 is still AC then the Schottke diodes or electrolytics are either in backwards or they are shot.

Check this out and let me know where you end up. I can take close up pictures of everything if you need to double check my instructions.

James, I hope this helps.

jameshuls · « **Reply #5 on:** 5 Dec 2008, 06:25 am »

Jim, I believe the problem is solved. I was measuring the voltage at C223 across the resistor, instead of from the earth to the component, so that is how I was getting 18 VDC. As it turns out, when I measured it correctly, there was 8.5 VDC going in (double what was coming out of the transformer - should there be 4.25 VAC as I am measuring?) and 0 VAC coming out, as you suggested. It looks as though a few of the resistors down the line got cooked as well, perhaps due to the excessive voltages not being dropped down by the R223? So, I'll order a few new ones (I now see the advantage of having spares around) and some new 10K caps and diodes for good measure. What benefits do you perceive with the FREDS vs these diodes? If you think it worth the $6 for the 4 (hard to go wrong really) I might as well do that, and then eliminate R223 altogether.

tubesforever · « **Reply #6 on:** 5 Dec 2008, 08:28 am »

I think Partsconnexion might have Fairchild Stealth hexfreds. Go for the 8A 1200v models as Bill Epstein has done. I selected 4A 600v models and I have 6.0v without the R223. I would prefer to have a little more voltage and use a Mills 12 watt wire wound to bring it down to 6.175 v at the H+

With the 10k caps, you need to make sure that the ground is properly oriented. Jim uses a square trace vs a round trace. That is cool. Make sure the ---- goes to ground side of the board.

For the Schottke diodes the silver band goes to the square hole. The silver band is demonstrating the cathode side of the diode.

If you substitute the Hexfreds you will get a smoother H+ supply. Jim Hagerman has stated here that the 10k ohm electrolytics are there to do the smoothing. His LC tank gets the H+ working just fine.

I like the hexfreds because the ring less. So the caps have less to do to filter noise. I figure this is a good thing. I sense a quieter and flatter background but this could be a placebo effect.

If you use hexfreds they are cheap. Make sure to shrink wrap them because they are conductive and will short out the circuit if the surfaces ever touch.

I am using the Panasonic TSHA caps for the three 10k LC tank. I would use the specified Nichicons if I had it all to do over again. I think they sounded just fine and were easier to fit on the board.

jameshuls · « **Reply #7 on:** 5 Dec 2008, 05:28 pm »

They are out of the Hexfreds, but what do you think of these in that position?

ON Semi MUR860 8A 600V FRED Diode; or,

IXYS 12A 1000V Ultrafast, Soft Recovery (FRED)
Epitaxial Diode (trr = 50 ns; TO-220AC package);

jameshuls · « **Reply #8 on:** 6 Dec 2008, 04:50 am »

Is my voltage of 4.25 VDC coming out of the transformer too high. From what I understand, it should be less than that. There is no question that the wiring is right, so if the voltage is wrong, then the transformer may be faulty, subsequently causing the problems down the line. I assume that an excessive voltage might fry the resistors downstream? Mr. Hagerman?

tubesforever · « **Reply #9 on:** 6 Dec 2008, 08:51 am »

This is a written technical description for everyone using 120V for their Cornet 2 connections. I think this is a good post for the archives regarding the incoming and transformer power connections.

The transformer windings on the Cornet 2 are designed for world wide conformance. The primary windings will define the secondary outputs and care should be taken to make sure the connections are exactly correct.

First we need to confirm that the IEC power cord plugging into the wall is properly connected to the power lugs on the PCB. In most 120v environments there will be three prongs on your wall outlets. There will be a wide blade, a round blade and a thin blade.

The wide blade will correspond to the neutral connection at the back of the PCB. The thin blade will correspond to the line connection at the back of the PCB. The round plug is the earth ground.

One recommendation I like to follow regarding the ground wiring, is to connect the IEC earth ground directly to the phono ground lug near the RCA's. Then I take this back to the PCB. That way the tonearm and/or turntable grounds are star grounding to your Cornet 2. This is very helpful in eliminating ground loop hums.

On the primary windings the wires connect to the lugs in this sequence. OO LL S and then S NN OO.

OO are simply dummy connections. the LL connections will be solid blue and solid black wires from the primary of the transformer. The two "S" leads are unused in the 120v circuit and nothing plugs into these leads. The NN leads will be solid brown and solid white.

If you have the Black/red wire connected to the LL by accident that could create some interesting voltages like what you are seeing on the board. The solid black and the black/red wire are identical except for a tiny thin red line that runs along the side of the wire. It is really tought to see. You must make sure this is connected correctly.

All the other wires unused wires should be connected to the OO on either side of the board. This is not critical since they are simply dummies.

For the H+ connections there is a Green/green-yellow/green connection. With my Cornet 2 the three wires here each supply 3 volts AC to the PCB. The Diodes job is to convert the AC to DC power. Diodes have ringing so Jim Hagerman also provides three 10kohm capacitors as an LC tank to smooth the dc supply. The 12 AX and 12 AU 7 tubes are designed for 6 volt heater supplys.

I get the best results at a voltage of 6.175 with my tube complements.

James, please double check the IEC connections to the main board and the primary connections to the PCB.

I bought my Hexfreds from PercyAudio.com. Digikey also sells hexfreds. 8 amps 1200 volts is what Bill Epstein is using. I am using the 4 amps 600v hexfreds with excellent results.

Let me know if the wall to IEC to PCB connections are correct. Then look at the primary winding connections to the PCB. If everything is good then we simply need to select the appropriate resistor at R223 to cut the H+ voltage to approximately 6-6.175 volts DC.

Cheers!

tubesforever · « **Reply #10 on:** 7 Dec 2008, 08:45 am »

Quote

Jim, I believe the problem is solved. I was measuring the voltage at C223 across the resistor, instead of from the earth to the component, so that is how I was getting 18 VDC. As it turns out, when I measured it correctly, there was 8.5 VDC going in (double what was coming out of the transformer - should there be 4.25 VAC as I am measuring?) and 0 VAC coming out, as you suggested.

[/color]

James,

Did you confirm 8.5 vac or 8.5 vdc at the top of R223? If you have vdc then you know the LC tank and Schottkes are doing their thing!

If you have 8.5 volts, you simply need enough resistance in R223 to bring the voltage down to between 6.0 and 6.2 volts for the tube heaters.

Since you have zero volts coming out, there is no heater voltages at all.

I would buy a series of wire wound Mills 12 watt resistors so you can roll these to your rectifier tube. You could buy sand resistors as specified in the original build but I find these to sound etched and grainy. Pun NOT intended. The sand resistors are not my sonic favorites.

Cheers!

jameshuls · « **Reply #11 on:** 7 Dec 2008, 07:18 pm »

It was 8.5 VDC, so the diodes and caps are doing their job. And to confirm 0 VDC (not AC as I mistakenly put below) coming out. The wiring is correct on the transformer, as I puzzled over the two black leads when I put it together, but when I turned up the lights, I noticed the red stripe on one. I had a kiwame 5W in there, but I will get some Mills 12 W. What values would you suggest?

tubesforever · « **Reply #12 on:** 7 Dec 2008, 07:57 pm »

James,

I originally purchased the specified Digikey sand resistor, a 1.8 ohm 5 watt piece. When I fired up with a Bendix 6106 Rectifier, I was seeing 5.86 on the heater voltage.

Some like to starve the tubes. It can take a thick sounding piece and make it sound more linear, however it also diminishes bass slam and dynamic swings. I used a Mills 12 watt 1.5 ohm piece and got the heater voltage to 6.175.

When I went to 4a 600v Fairchild Stealth Hexfreds, I ended up being able to just jumper the R223 all together. Bill Epstein decided to use the 8a 1200v Fairchilds but I am not sure he has his up and running right now. Has anyone heard from Bill????

I will purchase some 8a 1200v Fairchilds to spin in my Cornet 2. I would prefer the H+ to be about 6.175. That is where the Cornet 2 sounded best in my system.

I must be sick to have tried everything from 5.6 to 6.3 but heck....what else does a diy nut case do for fun?

This will be system dependent. So what worked out well for me might be less successful for other systems. I like to take all this "important" stuff to that level. Try a 2.0, a 1.8 and a 1.5 resistor and see what gets you closest to 6v. That is where the design center is for the 12xxxxx tubes.

I think the main resistor fried on you. I don't know if other resistors were effected, but if measuring a resistor on your board remember that if the resistor is in series with an adjacent cap, you might get fluctuating readings. When in doubt trust the resistor is ok and retest all the voltage test points with all the tubes plugged in and powered up.

Just for others reading in, if the voltages at the test points are off from their specified values by more than 10%, you can adjust the resistor values of the B+ drop downs to get that back to nominal. Email me if you need me to specify the locations. With my Mullard GZ rectifier tube, I have 392 volts at the B+ and even after the two stock drop downs my voltages were higher than I wanted to see. I just had to make minimal adjustments to the B+ step down resistors in order to get the voltages nominal.

Luckily this is easy and simple to accomplish. Jim's design work is superb - he uses a pair of drop downs probably to provide cascading protection to the overall circuit. The double PCB is really tough so you can solder in and remove pieces (carefully) and not have to worry that you will lift and trace and be stuck with running jumper wires.

ScottRT advised me to get a good soldering station before beginning the Hagerman projects and I am very thankful that he did. This was the best 77 dollars (including the tax) that I have ever spent!

Cheers!

jameshuls · « **Reply #13 on:** 8 Dec 2008, 04:16 am »

As always, thanks for the good advice Jim. I will let you know how it goes once I get my new parts. By the way, what are the B+ resistor step down positions? R104-108? I am going to be replacing some of these as well, as a number of these appear fried as well.

I have to agree on the good soldering iron. I got a digital soldering station meant specifically (High heat) for lead-free solder, such as the Mundorf gold/silver solder I used. I think it was only around $85 from Howard Electronic Supply and it has been fantastic!

tubesforever · « **Reply #14 on:** 8 Dec 2008, 08:11 am »

James,

The B+ comes down from the Rectifier tube and steps down with two 1R8 resistors. These are on the board as R220 and R221. My voltage was too high so I think I have a 2.2kohm and a 2.0kohm resistors in these spots. Like I stated earlier, you will not need a large swing of resistors. Just a few. If everyone just buys the 1.8kohm, 2.0kohm, 2.2kohm and 2.5Kohm resistors then you should have enough to mix and match to get the voltages right.

The H+ is handled with a single R223 resistor. I like to use the Mills 12 volt wire wound here because it will take whatever abuse the system throws its way without getting etchy sounding. I would get a selection of different values of these as well.

I hope the parts swaps gets things working well. If you go hexfreds make sure to email me. They need to be oriented properly to function. I have a post here somewhere about hexfreds with photos. The back silver side of the hexfred is conductive. It is the Cathode. You can tone this to the two legs of the hexfred to see which one is your cathode. The cathode goes to the banded marks on the PCB. You need to heat shrink these or you could get some nasty shorts or get zapped if you touch them with the power on. Heat shrink is our friend!

jameshuls · « **Reply #15 on:** 8 Dec 2008, 04:18 pm »

Thanks Jim. By the way, can you or anyone else tell me what is purpose of R104-108. Now that I am figuring out what each component is doing in the circuit, I want to know more!

ronpod · « **Reply #16 on:** 8 Dec 2008, 05:34 pm »

James,

There are others that know much more about the electronics that I but I will try to analyze this circiut to the best of my ability. Don't hold it against me if I royally screw it up.

R104-108 are in a circuit network that monitors B+ and H+. Resistors R105 and R106 forms a voltage ladder that feeds a voltage level dependent on H+ to the base of transistor Q100. Resistors R107 and R108 froms a similar voltage ladder that feeds a voltage level dependent on B+ to the base of transistor Q101. Remember that tubes are voltage devices and transistors are current devices; that is that when the transistor is turned "on" current flows through it. Resistor R104 connects H+ to the emitters of both Q100 and Q101. The difference or bias between the emitter and the base of these transistors are what turn them on. So when the Cornet 2 is first turned on, B+ causes Q101 to swith on current subsequently lighting the red LED. As H+ increases, Q100 turns on and Q101 turns off switch the green LED on and the red LED off. This all should help your expensive tubes last alot longer.

Hope others jump in to correct my analysis if it is faulty...
And I hope that this helps.
Ron.

jameshuls · « **Reply #17 on:** 8 Dec 2008, 06:14 pm »

So, if one or more of these resistors are fried (which they are) I might experience an LED that instantly goes green (as I do)?

ronpod · « **Reply #18 on:** 8 Dec 2008, 06:36 pm »

Is the fried resistor R107? (Just a guess) Yes, a fried resistor in this circuit could cause instant green. It is a good idea to replace all "fried" resistors.

jameshuls · « **Reply #19 on:** 8 Dec 2008, 07:37 pm »

Just going from memory (I'm at work) R106 and R108 (6.8K) for sure - they actually turned brown! And I believe 105 is as well. There may be more - I will have to check again. In any case, I am going to replace all of R104-108.