Handmade Electronics
Post Office Box 9114
Allentown, Pennsylvania 18105-9114

Updated:  July 23, 2010

Back To The Webstore

Capacitor and product quality bloviations
Updated: December 29, 2010

Magazine articles archive,Pentode SE operation


Murphy's Laws of Engineering (Originally: If anything can go wrong, it will):
1) The power transformer will always blow first in order to protect the fuse.
2) Paint will adhere to every surface but the one intended to be painted.
3) Self oscillating circuits won't.
4) Non oscillating circuits will.
5) An expert will always state the obvious.
6) Any wire cut to lenght will be too short.
7) A difficult task will be halted near completion by one tiny, previously insignificant detail.

Beware of the Baghdad Bob's of audio.

A little misinformation can go a long way.

When do I replace my tubes???
Here's 2 opinions from the companies who made them:
1) Genelex (MOV valve) of England- when a tube/valve reaches 50% of it's orginal transconductance.
2) RCA- when the tube stops working correctly (can it be any simplier than that?).

Practical troubleshooting & construction tips

You got it together....
...and it doesn't work (or barely works).

Here's a few ideas....

General: Before soldering, it is not a bad idea to clean all leads to be soldered with a little tuff of fine grade steel wool. Clean surfaces, free from oxidation make for better solder connections (especially helpful with low flux solders). Remember to clean up those loose strands of steel wool before power up!

Nothing happens
All connections soldered?
Did you correctly wire everything? Check wiring after an hour away from the unit. Make sure you didn't solder the wire insulation, and not the wire. Incorrect wiring is where most errors occur.
Is there power to the wall outlet?
Is the unit under test plugged into a wall outlet?
Is the fuse installed?
Internal AC wiring to the power transformer primary correct? Usually they're solid black leads.

Fuse blows on turn on
Fuse rating correct?
Unplug all tubes, does fuse still blow? Look for strands of wire shorting out filament or B+ a.c. line.
If, after removing all the tubes, and fuse doesn't blow again, replace rectifier only (observing filter cap voltages), and try again. Then add a driver tube, and another and then the power tubes until fuse blows again. At the point the fuse blows again, check that stage for faults.

If the fuse blows after rectifier warm up, and conduction, suspect the filter cap(s). While electrolytics may not fail in a catastrophic manner, they may develop leakage that will cause a high power drain. Get a VOM and take an ohm reading. While it has been suggested 1meg+ reading is ok, I suggest 2meg or greater resistance, or junk the cap.
Avoid old reformed caps, as they are just that, old. The electrolyte dries out over time, and that's where your heartache begins. Considering you may not be able to get a replacement power transformer cheaply, is it worth it? If you want to retain the original external appearance, bury new caps in the chassis, and leave the old ones in place, disconnected.

Disconnect all secondary leads from suspected transformer(s). Take an ohm reading from B+ pin on rectifier, and look for infinite resistance. If resistance reading is infinite, there are no circuit shorts. Suspect power transformer internal short, which calls for transformer replacement. Do this procedure to filament taps as well.

Low B+, or filament voltage
Correctly sized power transformer?
First stage filter cap connected?
Faulty rectifier tube

High B+
Many power transformers have a 115vac primary, which will cause a higher B+ and filament voltage condition, as many lines run at 120vac or greater now.
Tubes plugged in? The power transformer should be properly loaded first, before taking a reading.
Power transformer over rated for the job? Too much capacity may not allow proper loading.
Choke shorted? Take a resistance reading. If you're getting zero ohms, or infinite resistance, replace it.

Tubes do more than glow, they spark!
Faulty tubes? Check them out, best way is to use a known good set.
Tested and matched tubes are strongly suggested. Yes, you'll pay a bit more, but low prices may mean low quality tests. Remember, good work always costs more.

Tube plates glow bright orange.
Faulty tube? Replace with known good unit.
Check coupling caps (capacitors that connect to a tube's grid) for shorts. Infinite resistance is what is needed here.
Oscillations- usually faulty circuit design. In some cases, not enough B+ filtering.
Check for correct bias?
Contacts on tube sockets clean and tight?
Solder joints? Are they old, or embrittled? Refresh old solder joints by removing the old solder, and replacing with fresh.
Any lifted or disconnected traces or solder pads on printed circuit boards?
Properly matched output tubes in Push-Pull designs. Don't don't don't listen to anyone who says matching doesn't matter, It does!!! Don't be penny-wise and pound foolish.
Defective push-pull output transformer- both halves (center tap/B+ line as reference) should be within 10% of each other). Check each tap. Mis matched sections due to internal shorts can give the illusion of poorly matched output tubes.

Hum & Buzzes, Noises
1) RFI from light dimmers- try turning light dimmers off. Replace with low RFI types if need
2) Equipment next to any hum generating gadgets?
3) Power and output transformers 90 degrees to each other? Possible inductive coupling is happening (in some cases, 45 degrees may work better).
4) Chassis grounded to B+ negative line?
5) AC filament lines may require grounding on one side, usually through a capacitor (try 10uF  or greater).
6) Did you avoid multiple chassis ground points? Use "star" or "buss" grounding.
7) Do you have a bottom plate on your chassis? You may need this extra shielding.
8) Did you provide a good mechanical/electrical contact area between bottom plate and chassis? Provide a paint free surface at one point between chassis and bottom plate, and use a star washer to provide good electrical contact.
9) Is the B+ negative line, and filament negative line (on amps with dc filaments) connected?
10) Check for failing filter cap.
11) Vacuum tube rectifiers should not have heavy filtering, usually, the largest recommended capacitor is 40uF right off the rectifier. Some rectifiers, like the CV378, should have very low amounts, on the order of 4-8uF. Increase filtering after a choke, or a resistor. Solid State rectifiers can handle heavier filtering without stress, usually on the order of several hundred uF's.
12) Is there enough filtering to begin with? Choke input designs require heavier filtering (100-150uF minimum), AFTER the choke.
13) AC and DC lines bundled together? Not good....keep AC and DC separate, and, at 90 degrees to each other if possible. That style of wiring looks neater, but can introduce hum problems.
14) Did you do a hum balance adjustment on Single Ended designs?
15) Mis-matched output tubes in Push-Pull designs.
16) Possible output tube failure in Push-Pull designs.
17) Power transformer or choke buzzes when using choke input power supplies? Add a .47,.68 or 1uF capacitor off the rectifier. You'll retain most of the qualities of choke input with little voltage rise, and minimize that buzzing.
18) End bell screws on transformer or choke snug?

Crackling, pffff noises
1) Possible poor internal connection in capacitors, resistors, etc. Many cheaper brands use poor quality internal termination's, resulting in noise. Remember: Holco terminations do require care in handling. Also take into account component deratings in high temp environments.
2) Possible faulty rectifier tube. Try a known good replacement.
3) Possible faulty input or driver tube. Try a known good replacement.
4) Possible faulty output tube. Some current production types can not operate at or near book spec, and will start making noise when used either under high voltage and/or high current conditions. Recent tests suggest at this point in time, derate most KT88 or 6550 types to 32-38 watts maximum, keeping voltage at 525 maximum @ 60ma plate dissipation in Ultra linear mode (EI KT90 excepted). Chinese 2A3 double plate types sound quite good, but are prone to noisy operation. Newest production tubes show great promise.
The Chinese 2A3B and C single plate variants seem to work well.
5) Dirty tube pins or socket contacts.

Place equipment on a solid base.
Tubes run hot, try tube coolers or lower filament voltage slightly...6.1/6.2vac instead of 6.3vac.
High gain circuits will exhibit more noise.
6DJ8/6922/7308 types are just pain in the asses regarding microphonics. Avoid them.

Output tubes biased correctly?
Worn tubes. Try a set of known good ones.
Circuit design.
Rectifier section balance out by more than 20%.
Coupling caps too large in value for circuit design.
Power amplifier not up to driving speakers.

Other tidbits, product information:

Best electrolytic capacitor performance happens when the capacitor is run at 80% of it's voltage rating.

Not all small metal encased capacitors are oil filled. Some are tantalum, in many makers catalogs, the tantalums far out numbered the oil filled....so be careful about what you buy.

Kink in your solid core wire? Run it over the edge of a counter top, holding it fairly tight. After a few passes, kink will slowly work it's way out.

Use nail polish for general sealing. Good for wire where the insulation ends, or to repair the vinyl coating on Solen caps, as well as the epoxy seals on Solen's (which occasionally form small cracks when the lead is bent). A dab on screw threads will keep nuts on tight.

Make sure UX 4 pin type tubes...2A3,300B,811 are properly keyed before placing them in the socket. If you have to force a tube in, you may not have the two thicker pins in the right socket holes.

Check your current- Voltage to be measured divided by the resistance it must cross = your current reading.
Example- 8 volt cathode voltage divided by a cathode resistor of 820 ohms = .009756 amp/9.7ma.
In figuring plate dissipation, deduct cathode voltage reading.

.0001uF =.1nf
.0001uF = 100pf
.001uF=  1nf
.01uF=   10nf
.047uF= 47nf
.1uF=     100nf
.47uF=   470nf
.47uF=   470,000pf
1.0uf =   1,000nf
Personally, picofarads suited me just fine.

Calculator for parallel resistor values:
Computer torture page

1kilogram= 2.2 lbs.

Be prepared: You'll never know when you'll need these:

1 Furlong = 1.02960595 Leagues

1 Circular foot (US survey) = 0.0000073 hectare

Burning Questions Of Our times

Should a period be placed after the S in the name of President Harry S(.) Truman?

Tube type cross reference & substitution guide.
Tube Type No. Military No. European** Substitution type*,** Different pin-out(1)
0C3 VT200 N/A
1S4 VT210 DL91 N/A
2A3 VT95 N/A 5930
5U4G VT244 N/A 5AS4
5V4G VT206A GZ32 5AR4
5Y3GT VT197A U50,U52 5931 5Y4, 80
6J5GT VT94D L63 7193
6J7GT VT91A Z63 EF37,37A
6K6GT VT152 N/A 6F6/GT
6L6G VT115A EL37, KT66 1614,5881
6N7 VT96 N/A
6SK7 VT117 OSW3111 1635
6SK7GT VT117A DL91 6SS7
6SJ7M VT116 N/A 6SS7,5693
6SL7GT VT229 ECC35 5691,6188
6SN7GT VT231 ECC32 6BL7,5692
6U8,6U8A N/A ECF80 6BR8A,6FV8A
6V6 VT107 OSW3106 6K6
6V6GT VT107A
6X5GT VT126B
6Y6GT VT168 N/A
12AT7,12AT7A N/A ECC83,ECC810 6201,12AY7*,CV4024 6DT8
12AU7,12AU7A N/A ECC82,ECC802 5814,5963
12AX7,12AX7A(2) N/A ECC83,ECC803 12AD7,6681,7025
10 VT25
27 VT29
76 VT76 DL91 37
83 VT83 N/A
807 VT60 QV05-25
6550/6550A KT88/E.I. KT90
7119 E182CC 7044
KT88 CV5220 6550A/E.I. KT90
*- may not be optimum **- list is for parallel filament circuits only
(1)- If you're up to rewiring your sockets
(2)- 12BZ7 if heater supply can handle it

Back To The Store

-Since 1992-