Electron Engine ™
Printed Circuit Boards by Emissionlabs

EE40 Universal HV Power Supply board for rectifier tube, or silicon diodes.

Kit order Number: 311-040-39

MOST PARTS ARE INCLUDED. You receive:

Have you ever noticed, some amplifiers are totally free of hum, with a relative simple construction, while some others, build expensive, with huge capacitors and big effort, just keep on producing some residual hum? The answer, in order of difficulty is: 1) correct ground wiring. 2) The right circuit. 3) The right components. The answer is usually not: The capacitors are ere too small. More often the choke was too low inductance, too close to saturation, or even both. However the best choke, can not cure a bad circuit, and even a good circuit will hum if the ground wiring is wrong. EE40 connects right here.

Setting up correct ground wiring, is difficult to explain. In design books for engineers, that is usually a full chapter. Users who prefer to experiment until "it works", are hard to convince to do it another way. While those who have problems which they can not solve, are seldom helped by reading chapters of books. This is where the EE40 comes in. Just use it, and most of the possible mistakes are prevented. Besides it is a clean and compact PCB, preventing power supply wire mess in the amplifier. EE40 helps to find out the ideal circuit, but it may just as well be used as a final PCB.

There are quite a few cases where hum ist generated by unlucky wiring. Then as a 'solution' against the hum, capacitors are made too large, solving the problem not really, and cause other problems such as rectifier tube sparking. Also excessive capacitor charge peaks can radiate hum into the pre amplifier via air. The basic rule has always been: A low voltage power supply needs a large capacitor and a small choke, a high voltage power supply needs a small capacitor and a large choke. We encourage you to use the Lundahl high inductance, dual coil chokes (they have only such types), for which the PCB is prepared. There is no "must" to use such. Also a single coil chokes of your own choice can be used.

For the rest, please be not confused by the long texts below here, because EE40 is easy to use.

There are three ways to use EE40

1) Evaluation board on the bench. In this case, the tube socket is mounted together with all other parts at the components side.

2) Final board in the amplifier, with external tube socket. Connecting the tube socket with wires to the PCB, and mildly drill the heater wires. In this case, use the wiring numbers as printed on the PCB. This requires your own socket with solder lugs (not supplied). In this way, EE40 can be attached to the chassis at the place you prefer most. Consider the EE80 board to mount the rectifier socket in the chassis. This looks nice, and gives control of the rectifier function. For instance when you use the EE40 power supply loaded, like normal, the red LED of the EE80 board will be off. However when the power supply is used unloaded, by mistake, the red LED of EE80 will be on, indicating the output voltage is too high.

3) Final board in an amplifier, with the tube socket mounted on the PCB. In this case, the tube socket will be mounted on the BACK of the board, at the solder side. This is a little bit of a trick with the tube pin connections, but it is nicely possible, and it works well. For this, the positioning slot of the socket has to be mounted, with one pin rotated. This is prepared for you. For this, look at the SOLDER SIDE of the board, there is a clear white dot, near the socket pins. This dot, you have to align with the center slot of the tube socket. (This dot is only printed at version 1.0 or higher). The socket covers the text of pin numbering, where it is mounted. On the components side, interesting now, the pin numbering becomes visible, and the pin numbering are correct.

Please note: Close the solder Jumpers across R1 and R8, if you are not using those resistors. Otherwise the PCB will not work.

Features:

Rectifier configuration

 
Rectifiers type
Transformer has center tap?
Voltage doubling
Silicon rectifier used
Example for 250V AC. (DC voltage depends on load)
1
Silicon Only
Yes
No
D1, D2
250-0-250 primary will give appr 300V DC. R2=R3=31k. Take closest commercial resistor value.
2
Silicon Only
No
No
D1, D2, D3, D4
250V primary will give appr 300V DC. R2=R3=31k. Take closest commercial resistor value.
3
Silicon Only
Yes, but this wire is NOT connected to the board.
Yes
D1, D2, D3, D4
250-0-250 primary will give appr. 600V DC. R2=R3=125k. Take closest commercial resistor value.
4
TUBE Only
Yes
No
none
250-0-250 primary will give appr 300V DC. R2=R3=31k. Take closest commercial resistor value.
5
No
No
D1, D2
250V primary will give appr 300V DC. R2=R3=31k. Take closest commercial resistor value.
6
Yes, but this wire is NOT connected to the board.
Yes
D1, D2
250-0-250 primary will give appr. 600V DC. R2=R3=125k. Take closest commercial resistor value.

 

Filter configuration

 
Type
Choke Type
Wiring of connections
1-3 and 4-6
Note
1
C-L-C
Single Coil
Choke at 1-3, Wire Bridge from 4-6
This will keep coil at ground level for more safety.
2
C-L-C
Dual Coil

Lundahl choke, pin numbers equal 1-3 and 4-6.

If not Lundahl: 1 and 6 have same polarity

Lowest noise
3
C-R-C
  Resistor from 1-3 and from 6-4. External Resistor, with cooling. Has residual AC. Possible for Push Pull with feedback.
4
L-C
Dual Coil

Lundahl choke, pin numbers equal 1-3 and 4-6.

If not Lundahl: 1 and 6 have same polarity

Leave away: C1, C2, but still use R2, R3

Ultimate way, but requires 2...3x larger inductance. Allows highest DC power to be pulled from the rectifier.
5
L-C
Single Coil

Wire Bridge from 6-4.

Choke at 1-3.

Leave away: C1, C2, but still use R2, R3

Ultimate way, but requires 2...3x larger inductance. Allows highest DC power to be pulled from the rectifier.

 

BHC Aerovox CapacitorsCapacitors C3 and C4 are included. These are 105°C Type.

Please note this: These capacitors are superb quality NOS Audio Capacitors from 1995, made in England, but they should be formatted before first use. This is nothing but a very slow start, for the first time. The EE40 board itself can be used for this as follows: When building it, do not solder the capacitors on it yet and do not connect the choke. Now the connections 6 and 3 supply an unstabilized voltage. Connect 6 to the plus of one capacitor via a 1 Meg resistor, and 3 to minus, and hook up a DC voltmeter to the capacitor as well. Switch on the EE40, and this will slowly charge the capacitor. Monitor it, and stop at 350V, or as far as the EE40 goes. Then switch the EE40 board off, and wait for the capacitor to be discharged, before removing it. Then repeat with the other capacitor. Even so, you can repeat the formatting just by switching on and off the EE40, and you will see, the second time it charges much faster. When the capacitor charges in 7...10 minutes to 2/3 of the final voltage, the formatting was successful. And it always will be. Though not strictly needed, re formatting is ideal for never used before capacitors. After that, capacitance is closer to the rated value, and less leakage, and the positive effect will stay.

Function of the links across R1 and R8.

These links should initially be closed, to get the circuit to work. But later you should decide on the best value for R1 and R8.

R1 is used to adapt the transformer voltage to the tube. The actual voltage can be measured with probes, across the test points "Test Fil". Note, tubes can tolerate only maximum 5%. If the heater voltage is too high or too low, the rectifier may spark at a cold start.

R8 adapts the transformer Raa specification to what the tube needs a a MINIMUM. So when the tube data sheet writes Raa>130 Ohms, and the transformer Raa is 70Ohms, R8 becomes minimum 60Ohms. If Raa of the transformer is too low, the rectifier may spark at a cold start.

In reverse, if the rectifier does not spark, it does not mean R1 of R8 are correct, or not important to look at. Most early rectifier failures are caused by this.

Use of EE40 with some Lundahl transformers

Lundahl transformer
Transformer Center Tap
Voltage doubling
Artificial Center Tap
DC Output voltage unloaded.
Maximum Load.
HV windings in series. Connect 6 to 8. There no Center Tap. "CT" of PCB stays empty
No
Yes. Insert D3 and D4. 490V. See also Notes on this.
630mA
HV windings in series. Connect 6 to 8. There no Center Tap. "CT" of PCB stays empty
No
Yes. Insert D3 and D4. 322V. See also Notes on this.
1A
HV windings in series. Connect 6 to 8. There no Center Tap. "CT" of PCB stays empty
No
Yes. Insert D3 and D4. 490V. See also Notes on this.
630mA
HV windings in series. Connect 6 to 8. There no Center Tap. "CT" of PCB stays empty
No
Yes. Insert D3 and D4. 700V. See also Notes on this.
43mA
Unloaded voltage would be 743V, this is not possible with EE40.
Connect "CT" of PCB to 22+24 of transformer
No
No. D3 and D4 are REMOVED. 350V. See also Notes on this.
160mA
HV windings in parallel. There is no Center Tap. "CT" of PCB stays empty
No
Yes. Insert D3 and D4. 350V. See also Notes on this.
160mA
HV windings in series. There is no Center Tap. "CT" of PCB stays empty
Yes
Yes. Insert D3 and D4. 700V. See also Notes on this.
80mA
Connect "CT" of PCB to 22+24 of transformer
No
No. D3 and D4 are REMOVED. 350V. See also Notes on this.
260mA
HV windings in parallel. There is no Center Tap. "CT" of PCB stays empty
No
Yes. Insert D3 and D4. 350V. See also Notes on this.
260mA
HV windings in series. There is no Center Tap. "CT" of PCB stays empty
Yes
Yes. Insert D3 and D4. 700V. See also Notes on this.
130mA
HV windings in series. Connect 19 to 22, this is the Center Tap, connected to "CT" of PCB
No
No. D3 and D4 are REMOVED. 490V. See also Notes on this.
200mA
Voltage doubling would give 980V unloaded, but PCB and Capacitors are not made for this. Not possible with EE40.

 

 



NOTES and some tech talk.

The strange position of Fuse F1. Before we dig into the virtual center tap, here is why the fuse is at this unusual place. Few people choose this position, but I think there no better place. Most of the time, the fuse is directly placed at the output, but no normal radio fuse can fuse 700 DC (which is the unloaded voltage) safely. If activated, the fuse becomes totally black inside, but there is a risk it may explode so violently, it is completely gone. Even the two metal parts may not be at it's place any more. Reason is, at high DC voltage, plasma may build up inside the glass fuse. This is a conductive gas. That makes the fuse explode. Sand filled fuses are safer, but I wouldn't bet on what happens if you fuse 700VDC with a 230V AC fuse. That is why we put the fuse in the center tap, which is grounded, and it is at AC level instead of DC. An additional advantage is now, the fuse is at ground level. It's much safer. We do not sell fuses, but when you buy one, make sure it can fuse the AC voltage of the transformer high voltage winding.

The Artificial center tap. Please read also the "fuse" part. above here. A tube has only two diodes, and it is not a bridge rectifier. To use a tube rectifier, requires normally a center tapped HV winding. If this is missing on the transformer, we can still use this transformer, when we create an artificial center tap, with two silicon diodes. Before explaining this, first look at a classic center tapped circuit: The center tap is GROUNDED. This is very important to keep in mind. To understand how the VIRTUAL center tap works, look at the negative of C2 in this schematic This is ground level. We have the choke in between, but it filters out only residual AC before the choke. The virtual center tap would work just as well without the choke. With a single coil choke from 4-6 and a wire link from 1-3, C2 is even hard-grounded. And yes, feel free to try this with the EE40. So regard the negative of C2 as GROUND level, just for better understanding.

The fuse F1 plays no role for the functioning of the circuit. To the fuse is D3 and D4 connected, and the transformer has no center tap. Here is how it works: Since we have AC signal, ALWAYS either D3 or D4 is in conducting state. This means, in conducting state there is only 0.7V DC across it. This means at the positive half of the mains, D3 is conducting and the fuse is clamped to ground by the conducting diode. At the other half of the mains, D4 is conducting. So all of the time, the fuse is clamped to ground. The only difference with a real center tap is this 0.7V DC of the conducting silicon diode, but in compare to the voltage drop across the tube rectifier (some 20...35V) this can be neglected. This doesn't change the rectifier tube curve, other than by shifting it up 0.7VDC. As the tube pumps into 400V DC (or so) of C1+C2, this 0.7V plays no role at all. So with the diodes, the tube pumps current into 400.7V and otherwise into 400V. The tube itself has a resistive character, and a current limiting effect for peak current, which is why we use it. This will not be changed in any way. Virtues of tube rectification (less aggressive charge spikes, slow start and current limiting) it stays all unchanged. This way, indeed still normal tube characteristics will stay.

Voltage doubling. Another thing we can do with the artificial center tap, is voltage doubling. A 250V-0-250V winding with tube rectification would normally give 300V DC (depending on the load of course). WIth an artificial center tap, voltage doubling occurs, and we get 600V DC. Note: In case of voltage doubling, do NOT connect the center tap together with the diodes D3 +D4. So the center tap must not be connected to the PCB, but D3+D4 must be soldered in, and then you have voltage doubling. Also with a tube rectifier. Correct connections for this, are in the table..
What is a center tapped winding? Under cost pressure, transformer winding companies over simplify this, unless you give clear orders how to do it. Here is why. Because of AC, and because we use rectifiers, a center tapped winding carries only current through one section of course. This would be fine with a toroid, or with any other core, when BOTH halves of the tapped winding are wound TOGETHER, so physically at the same place, and in addition have the same geometry too. This can only be done with bifilar winding, or a split chamber. Bifilar is not possible with High Voltage. A split chamber, you can check if it there, from the outside. Is this not visible, then you don't have it. The cheaper way, the usual way, to lay one package over the other, will lead to uneven wire length, and uneven load, due to copper resistance difference, and uneven magnetic coupling. Effectively this will give some l DC magnetization of the core. Then, at large voltage, this would saturate the core, which gives a specific noise, which you may have heard before. It sounds like "hummmm". Particularly under dimensioned transformers will have no air gap, because air gapped cores cost more, and are larger dimension, even requiring a longer copper wire. Meaning however, under dimensioned transformer can withstand absolutely no DC component. To be on the safe side, such windings should be dimensioned a factor 1,7 larger, is what Per Lundahl told me. Will EL CHEAPO company do this? No, I don't think so. Simply inquire if their center tapped winding balanced (symmetrical) and if the transformer has an air gap. When they dispute the need, or don't understand why you say this, you can already expect what you are getting. Alternatively, for low cost transformers, you should think of the HYBRID connection. This avoids the problems resulting from a not correct constructed tapped winding.

Raa Winding resistance. Rectifier tubes require always a minimum Raa windings resistance. This special knowledge of tube transformer making has gone with the wind of modern times. But for the tubes itself, this requirement has not gone at all. Each rectifier requires a certain MINIMUM windings resistance, or the tube will have low lifetime, or even show a white spark at switching on. In classic circuits, transformer Raa is measured at the tube socket, between the tube Anodes, when the amplifier is off, mains cable and rectifier tube unplugged, and capacitors discharged. In many cases this value is much too low. This major design error can be fixed by adding an external resistor in each side of the Raa winding. This resistor has half the value of the total Raa required. If a tube needs Raa of 250Ω Ohms, and you measure 100Ω only, there is 150Ω missing. So to add this 150 Ohms, you need to serialize two resistors of 75 Ohms at each transformer winding ends. Alternatively you can add one single resistance of 75 Ohms at the Cathode Tap, which has the same effect. Then, from the tube's view you have Raa of 250 Ohms again. So to be sure, unplug the mains, wait for capacitors to be empty, unplug the rectifier tube, and measure transformer Raa at Pin 4 and 6 of the Octal socket. This value is important, it may not be too low, or tube problems can occur, and transformer hum will be too loud also. This simple measurement is HIGHLY interesting to do on expensive tube equipment. Any yes, chances on a mistake is higher than finding no mistake. It is sad but true, and it caused so many rectifier to die too soon. But really, we can not have transformer builders chance the requirements of tube data sheets?! That would be silly.
Loaded voltage depends on output current, but it can be lowered, by using very small values FOIL capacitors at C1 and C2, like 1...5uF. These quite some AC current and should be capable of it. Mainly that will be such with large dimensions. Otherwise they have too much internal resistance, and get warm. Or even leave C1 and C2 away completely. This will be on the cost of more AC ripple, which increases with the load. So this makes mainly sense at low current, and you find the loaded voltage is higher as expected. Another way to lower output voltage, is to add series resistors in the HV winding, there will be AC ripple across this resistor, which has two effects. First output voltage will drop, and second, t this AC voltage gets removed from C1 and C2. So filtering will improve also. As a third method you could combine both. So get the initial drop with a smaller capacitor for C1 and C2, and then make fine adjustment with a series resistor. The Resistor for this is R8. This resistor will become hot, and must be an external power resistor. So there are only wire connections for this on the PCB. If not used, the solder bridge must be closed.

The reason why dual coil chokes are made. Look at the schematic, and you will see how a dual choke cuts the ground path. That is the whole reason for it. The below schematics explain the effect, but is stays difficult to understand immediately

  • Here is the usual schematic1 and schematic2, which also applies for any single coil choke.
  • Schematic #3 applies for a dual coil choke. The AC ground path is cut, because for 50Hz, the choke is very high resistance.

 

FAULT FINDING

The fuse F2 blows often. The mains fuse F2 should be slow type, and not chosen to critical. Depending on the transformer and the load, start up current can sometimes be quite high. The mains fuse is only for emergency and can be chosen higher as the steady state current. The normal fuse for overload is F1. If the fuse F1 blows too often, this means the choke is too fragile for the job. You can recognise such already by too small dimensions. WIth an oscilloscope, the start up current through the choke can be made visible. Ground it to the PCB, and connect the probe to the 'TEST-Minus' test point. 1 Volt equals 1 Ampere. This should be close to a DC current, since such an inductor can not pass AC current.. Any nasty AC peaks, indicate saturation, which may blow F2.
If the fuse F1 blows too often, this means the choke is too fragile for the job. You can recognise such already by too small dimensions. WIth an oscilloscope, the start up current through the choke can be made visible. Ground it to the PCB, and connect the probe to the 'TEST-Minus' test point. 1 Volt equals 1 Ampere. This should be close to a DC current, since such an inductor can not pass AC current.. Any nasty AC peaks, indicate saturation, which may blow F1. If you have no oscilloscope, use a voltmeter to measure across the test points. Check on AC and DC.
Rectifier tube heater is not glowing. When there is AC voltage at F1 and F2, but the tube is not glowing, either apply the Link J1 if the transformer is made for the tube, or use a resistor R1, to adapt the heater voltage. If no solder Link J1, and no resistor R1, the tube can not work.
Rectifier tube heater is glowing but there is no high voltage. So the blue LED is off. Check if Fuse F1 is good. Was the link J2 inserted? Or otherwise was R6 inserted? If no solder Link J2, and no resistor R6, the tube can not work.
The blue LED burns, but there is no DC at the output. The PCB fuse was blown, due to a short.

A lot of hum on the output signal. If using a Lundahl dual coil choke, you may have connected one of the windings in reverse. AC hum should be almost none. So when there 20V or so, hum signal, this is the reason. With another brand choke, sometimes they are tiny dimensions, but there is no magic. It may by over specified with DC current, but also AC VOLTAGE across a choke is limited, and you need to MEASURE what AC you have across the choke, and compare this to the specification.