PART 2) . The list of amplifiers which need special attention

The following table, is about some personal experience, how to best deal with this model amplifier. We write it here, to make sure you use the products in the best possible way.

Air Tight ATM-1
CARY Audio
Kondo Japan): Model KSL KEGON
Kondo (Japan): Model NERO
Manley 300B Preamplifier 'NEO-CLASSIC'
Mastersound 300B-SE
Mingda Meixing
Tektron TK2A3-50M. Part 1.
Tektron TK2A3-50M. Part 2.
VAC 30-30
VTL 185
VTL. Model: ST 150
Welborne Apollo





The schematics were not available.

I was given a hand made schematic by a person, who helped himself. Imagine, he is not the only one with that problem, but perhaps the only one that draws the schematic before he begins. Later I was given this Schematic here

From this schematic it gets clearer what you need to look at. There is a phase reversal stage in here, with two ECC82, and it has an adjustment potentiometer. Adjusting this stage is 'overall' so it also affects the EL34 stage, which is drives. I would say, the best way is adjust this with an oscilloscope, so the El34 output stage is driven nicely symmetrical at full power. You need to do so, after bias adjustment. Now comes, what I think is important to know. These ECC82 tubes are Electro Harmonics tubes with the Air Tight logo stamped on it as well. Generally speaking, Electro Harmonics I have found to have a much larger parameter variation than NOS tubes, and the EHX ECC82 seems to draw less current than an average ECC82. So I would say, EHX is not an ideal tube for this, but it can be used. However, it was chosen, and since the amplifier is adjustable it doesn't matter anymore. Since the adjustment of the amplifier is an adjustment for the individual set of tubes, the best way is to have the amplifier re-adjusted by a real specialist after a tube change. Furthermore, since two random ECC82 can be quite different, you should never exchange those amongst each other by mistake. Needless to say, this can happen quickly! So the recommendation is, when you have those ECC82 replaced, you should always take what is called a balanced, matched pair.

Something similar applies for the EL34. Each channel (Left and Right) has it's own balancing potentiometer, so in the end you'll end up with different settings for those two potentiometers. Once that is the case, you cannot exchange the EL34 amongst each other. So the moment you pull them all out, you should mark them carefully 1....4.

This said, it is obvious, you cannot just replace the ECC82 and EL34 by 'some' tubes. The EL34 (Left and Right Channel, are controlled by the SAME bias voltage, so for that reason you can only put a matched QUAD in this amplifiers, and not two pairs.

Overall this is a really very good schematic, and once adjusted correctly, the EL34 stages should work very good. Apart from what others say or do, my opinion is, best results will only come from a proper adjustment by using an oscilloscope.

Disclaimer: This schematic was hand drawn from an ATM-1 and it may have mistakes in it. So no guarantee for the drawing being correct.


Sometimes a tube can violently spark inside the bulb.

This amplifier is DC coupled. With such concepts, you always need to understand which tubes warms up first, and what this does to the bias of the next tube. A good designer make a worst case design. Like try things with a very weak, but 'just ok' 12AT7, and a 2A3 of which he knows it is a very strong one. That will sure 'spark' a surprise sometimes, since the bias voltage of the 2A3 depends on the plate current of the 12AT7. What happens at warm up, the 12AT7 tube heats up slower than the 2A3, this cannot be changed. The result is, the grid voltage of the 2A3 for a few seconds higher than normal, totally undefined, and even POSITIVE in many cases. So you are only lucky if your 2A3 is not working yet at that moment. Most of the time though, the 2A3 will already be able to pull current. The better quality and the newer the tube is, the more current. Some tubes will draw so excessively much current at POSITIVE grid voltage, you blow off a chip from the filament, and the tube will 'fire' with a white spark.

1) Amplifier Version with without the soft start.
  • A white spark is always (no exception) an amplifier issue. So the problem we see here, is not related to 'EML tubes only' though we heard some amateurs say this, because 'El Cheapo' tubes seldom fire. Even so, tubes that do not fire, still a suffer a lot. I would say a NOS Western Electric probably will fire as well. Even some vintage Osram and an RCA datasheets mention the 'firing' as an amplifier problem. If you fire a Western Electric 300B, it is just as bad for this tube as for an EML.
  • A positive note, this firing is only bad when you repeat it too often with directly heated tubes.
2) Version with soft start option.

This basically solves the spark issue with this amplifier, and this is a good, and mature product. However, if you want to switch the amplifier off an on again, you need to wait 20 minutes in between (or DAMAGE will occur). Don't forget this! Actually this is normal with many products. The only thing is, here you need to wait 20 minutes which is a long time, when you are waiting.

3) Heater voltage of 2A3 tubes.

It seems the heater voltage of the paramount is too low. 2A3 must run on 2.5 V heater, tolerance on that is 5% for all brands 2A3. The heater voltage was reported 2.32V on the chassis inside, by one of our customers. This is by itself almost 8% off, and already below minimum. Moreover, because of 50mV voltage drop over each heater pin contact, there is another 0.1V drop. So with 2.32V at the socket solder connections, there is only 2.22V on the tube pins itself. It is recommended to have 2.6 Volts at the tube socket solder connections. To correct this, you can add an electrolytic capacitor, directly on the tube socket heater pins. Here is the text from Bottlehead company: 'This capacitor should be rated at 6.3V or more, and sized around 4,700uF to 10,000uF'. We recommend 10.000 uF.

  • Guarantee on EML tubes if used in the Paramount without soft start option:
    You need the upgrade kit for this, to add the soft start. (Also read the next line).
  • Guarantee on EML tubes if used in the Paramount with the heater voltage below tolerance:
    No. (see above text). In case of a tube problem due to under heating, please contact us by email. If EML, chances to repair them are high.
  • Guarantee on EML tubes if used in the Paramount with soft start option + correct heater voltage: :
    Yes. (Thanks Paul Joppa from Bottlehead for the good support)

Note from EML: We recommend to retrofit all paramount with the Soft start kit, and also adjust the heater voltage of the 2A3 such that you have 2.6V DC on the solder connections of the tube socket, for the tube filament.

General link to EML factory page about what causes a white spark: Read more here and here


Very good and much recommended amplifiers.

They push the 300B or 2A3 tubes close to the limits, which is of course possible, but you need to be aware of what it does to the tubes, and lifetime in general.

We recommend following:
  • Use EML 300B-XLS for CARY amplifiers using 300B as output tube.
  • Use EML 2A3-S for for CARY amplifiers using 2A3 as output tube.
  • If you want to use EML 2A3-mesh, you have to use those that are selected by us for a grid voltage not more negative than -47V.(So a tube of -48V should not be used). This is the voltage that is written on the outside of the boxes. Please ask for such a tube, and you are fine with the CARY. We supply this service to you at no extra cost.
Kondo (Japan): Model KSL KEGON

The issue is, the Parallel Single Ended output Tubes are hard wired in parallel, without a way to check and adjust the bias for each tube individually. This is very unwise to do, it assumes things about tubes which are not realistic.

This amplifier hard-wires two 300B tubes in parallel. (see schematic) So, without any precautions whatsoever, the terminals of the 300B tubes are connected in parallel, and that's it. This is dangerous with random tubes. Since this method was used here, we need to see how we deal with this. It means, this amplifier will only work safe on a pair 300B tubes with special selection methods, and definitely cannot work on unselected 300B tubes, and also not on just any matched pair. What you need here is EML 300B tubes with factory tested grid voltage from -56V.... -60V, by the Emission Labs test method.

So how is the Adjustment done? Well don't ask us, ask the manufacturer! From hear-say we expect you could measure the DC voltage over the Output transformer primary, and derive the plate current from that.

We give only guarantee when you use EML tubes like this. Another method, or another brand, it means guarantee will void. So please take care of the above, and you will love the results and the wonderful sound of this amplifier!

A SERIOUS WARNING I have seen an 1991 circuit diagram without bleeder resistors on the High Voltage Caps. (!!!!!). This is extremely dangerous when you open the amp, even when 'off' the power supply caps can be charged still for hours or days. This is unlawful in all countries.

All together this amplifier is regarded a potential problem maker. It should be serviced only by experienced people, and it needs special help tools to check the bias of each tube INDIVIDUALLY. We talk here about the JAPANESE KONDO, older version.

Kondo (Japan): Model NERO

This amplifier hard-wires two tubes in parallel. It is generally the wrong to do this. We need to deal with some difficulties. It means, this amplifier will only work safe on a pair 2A3 tubes with special selection methods, and definitely cannot work on unselected 2A3 tubes, and also not on just any selected quad, of which the seller says they are matched very tight, but in reality he did not measure grid voltage at the NERO settings. So ask him for these values of the table below here, and if he cannot give you this, these tubes are not what you want. IN THE NERO, you need a quad of matched EML 2A3 tubes with specially tested grid voltage. First, you need to measure the grid voltage of the tubes, with a voltmeter. Voltage is measured against ground. This is in the working amplifier, with the old tubes inserted, presuming they still work. If so, do the measurement and let us know. For legal reasons we must say: Do such a measurement only when you are sure you are qualified for it. If not, let a good technician do this for you.

We use the following table:
Tube grid voltage tested with DC Heating.
Amplifier setting
40,75... 41,75
41,25... 42,25
41,75 ... 42,75
42,25 ... 43,25

42,75 ... 43,75

43,25 ... 44,25
43,75 ... 44,75
43,75 ... 45,25
44,75 ... 45,75
45,25 ... 46,25

There are a few strategies to replace your tubes:
  1. Plug & play may work by coincidence, or damage the new tubes slowly if they are biased too high. This cannot he heard or seen, until damage takes place.
  2. Measure the amplifier grid voltage, so we can send just the right tubes. This is a good way. You don't need to change the amplifier settings, and also you can swap with your old tubes too if you like.
  3. If the amplifier settings are outside of the of range of this table, it means somebody adjusted it for a curious set of tubes. In this case you can order a quad of normal, good tubes with values from this table, and you have to re-adjust the amplifier grid voltage for these values. After the adjustment, you cannot use your old tubes anymore, only the new tubes.4) You want to adjust the amplifier for unknown tubes. This requires a measurement of the plate current, which is not provided in this amplifier in a logical and normal way. It can be done, but it is no do it yourself job. After the adjustment, you cannot use your old tubes anymore, only the new tubes.5) If you don't have the technical skills, or just trying to learn this, you should not practice this on your own. Bring the amplifier to a tube technician, and give him this procedure. For the NERO we recommend EML2A3-S because this tube is very forgiving with accidental overload. If you want to replace the tubes by NOS or by EML 2A3-Mesh, you need to make sure you understand the above procedure very well.

We give only guarantee when you use tubes under the above conditions. Otherwise, guarantee will void.

So please take care of the above, and you will love the results and the wonderful sound of this amplifier!

Manley 300B Preamplifier 'NEO-CLASSIC'

this is not a complaint or a problem about this very nice pre amplifier. It is just we want to point out here, you cannot operate this amplifier with EML 300B-Mesh tubes. The mesh tubes are recommended to bias at Maximum 28 Watt, and it seems this amplifier runs the tubes above. It is a bit high for a pre-amp, but fair enough this is Manley's choice. So we recommend normal EML 300B, and since this seems an pre amplifier on 'steroids' also the 300B-XLS is a good candidate.

Mastersound 300B-SE

this is one of the best 300B amplifiers I know. They are build around a decades old, very mature schematic, and the output transformers are great quality. The later types have electronic regulation for the heater voltage, making this amplifier very universal. This deals better with mains voltage variation, and allows other tubes than classical 300B only. However, they do need tube adjustment when you replace tubes. Failing to do the adjustment, may cause problems. You can check here, for some more information, about tube exchange.

Mingda Meixing

this amplifier loads the 274B tube with an insane value of 100uF capacitor. This is absolutely not allowed by any tube datasheet. The 274B datasheet says maximum 4uF. Then, after the choke comes ANOTHER 470uf, and after this via a 2k resistor ANOTHER 470uF. This schematic was not made by using the tube datasheet. Meaning it was made by trying it out. At any problems, the designers will repeat as always ' come we never had a problem with it...' This can cause the tube to spark, and sparking or not sparking, this may cause tube defects or later. You can check the amplifier's schematic at, this is an external link.


Issue: Some of their amplifiers use these metal rings which are positioned too high. We have seen socket contact problems result from this, as tubes do not go inside as deep as they they should.

'Just not' glass to metal contact
Here you see glass to metal contact

Also the glass contact to the chassis is a potential problem. The customer who owns this amplifier contacted the manufacturer, and the reply was, there is no support. Reason is, as they say, there is only one reference tube, and that is Western Electric.

This answer was typical, and can not be accepted, because Western Electric never defined the diameter of the glass, at the point where it comes out of the tube base.

When you let the glass disappear into the amplifier chassis, we say this is wrong to do, for power tubes. Even when they fit in, this can also cause failures with Western Electric tube. It heats up the socket from the inside, and it can cause problems.

So Silbatone offered no solution. On my advise, the customer solved it himself in less than five minutes. For this he simply removed the 5mm distance holder, that you will find at the inside of the amplifier. This will bring up the whole tube socket 5mm higher. The ring can stay, and the problem was gone.

That solved the whole problem afterwards. Why do amplifier manufacturers react so arrogant for no reason, that we have to write it here?

We say here, real GLASS CONTACT or almost GLASS CONTACT to the chassis is definitely a problem. This amplifier builder may seem no problem initially, but if there is mechanical stress on glass, it can crack spontaneously after years. And yes, also with Western Electric, because their glass is also made of glass.

Tektron TK2A3-50M. Part 1.

July 2020.

I had a case of selling two mint condition NOS 2A3 by RCA, which tubes both failed after 3 months of occasional use. Before sending, all data was 100%, and it was a perfectly matched pair as well. The failure mode was a bright red glowing anode, and and the amplifier stopped working. So the angry buyer send them back as 'Faulty tubes'. However I could diagnose the burned grids with both tubes. In any case this is an amplifier failure. Tubes don't burn the grids by themself. Moreover in case we do have an exceptional failure, that is always related to just one tube. Because that is what exceptional means. However when two tubes of excellent reputation show in very short time the same defect, with 98% probability we already talking about an amplifier problem. They were tested for grid current, and there was none. So the question is only how they got over heated. However the buyer was over asked with this, he can only 'roll tubes', and does not know what grid current is, or what it does.

The tubes in there are used with fixed bias. When looking on the RCA data sheets from the 1930's they indicate greater care is needed with fixed bias than with auto bias. The most important thing they write here, is not exceed the grid resistor of 10k or 50k (depending on the build year of the tubes). So any lower value is possible, but not a higher value. If you do, it will initially work. However slowly the tube grids will develop some small amount of grid current, which current has no other path than the 50k resistor. This will cause a mall DC voltage, which again will increase the plate current of the tubes. The result is more heat, and the grid current will grow further. This may reach an instable point, and suddenly heat begins to avalanche, and the anode glows red hot. At this moment, the tube self-destroys, because Barium ions from the cathode gets freed due to the excessive heat. Since these are positively charged, they are attracted to the grid. This will cause grid emission, which effectively is grid current. This current comes "out" of the grid, and causes a DC voltage across the grid resistor. And now problems begin. This voltage will offset the bias very much, and the tube will draw 150% or 200% of normal current. The tubes will glow bright red, and plate current becomes 400% or more. So the fuse will blow. After cooling down, the tubes initially seem to work again, because the grid has cooled down, and like cathode emission needs heat, grid emission needs that also. After a few minutes that temperature will be reached, and the process starts from the beginning: Red plates again, and a blown fuse again.

A burned grid, is like the chicken or the egg, who was first. You just must prevent if from happening the first time. This is why a value of 10k or 50k is indicated in the RCA data sheet as an absolute maximum, and you should not regard this 'defect' limit as a safe value. I asked Tektron, and they confirmed they used 47k. Which is pretty close to 50k. According to a July 2020 email I have from Tektron, such problems occurred before, with NOS 2A3 and NOS 50 tubes, but it is a tube problem, they say.

So, a tube problem. Well I see it not like that. To my opinion, 47k is too close to the fatal limit of 50k, and it exceeds the RCA recommended value of 10k. 2A3 amplifier articleHere is a copy of a most excellent article in RADIONEWS from June 1933, writing in detail about how to design a Push Pull amplifier with a great new tube from RCA, called the 2A3. It is on page 721. On this page you can read this:

Quoted text from 1933 article. 'With fixed bias, this resistance should not be more than 10,000 ohms. Higher grid resistors may cause the grid to lose bias due to grid- current and the plate current will then rise so high as to damage the tube' Please note, this is just a text from an author in this (leading) magazine. In the end what counts in RCA documentation, and this is here:

2A3 datasheet with 10k (Until Manual 12)
2A3 datasheet with 50k (Manual 13 and higher)

So as from the printing date of Manual 13, the tubes can be used with 50k. However.... HOW CAN YOU BE SURE the user takes tubes, made after that date? This is impossible to say.

So this point needs A LOT OF ATTENTION when NOS tubes die from a burned grid, in this amplifier. My opinion is, you can not blame the user or the seller of the tubes for this. This is a plain question of the amplifier itself.

So where are we now? For the use of 50k grid resistor, you can take the risk for yourself if you think it is worth it, but I fail to see WHY 50k is so important to take? 10k resistors are not more expensive.

I see this over and over again, when people are reading data sheets. When it writes there a maximum value for several parameters, people consistently confuse this with recommended values.

So my recommendation for NOS is MAXIMUM 10k.

About the use of EML tubes with 50k grid resistor: These are an exception. Unlike with NOS tubes, with EML, 50k is no safety risk. EML grids are physically much larger and do not get very warm anyway. Unlike the RCA grid which is known to be fragile, and can suffer a thermal overload quickly, as we have learned with this amplifier.

Tektron TK2A3-50M. Part 2.

October 2020.

Now it gets interesting! Another Tektron with damaged tubes. This users bought new 45-Mesh tubes from Emissionlabs. We build those ourself, and I know what is inside. They can not have grid emission, because they're made such that this can not happen. But AGAIN a customer who reports red-glowing anodes. So it was immediately clear we have a bias problem here, but what what problem can that be....? The buyer already contacted two dealers, and myself also. Then he send me pictures of the problem. And that explained it.

Tektron has a manual saying you need to set the 45 tube to 30mA. And that is exactly what the customer did, he wrote me so.

Here is what he send me. He send it for 30mA exactly.

I did not question the setting, because I know the Tektron has meters. But hold on... how can the anodes glow red at 30mA? For RCA tubes that needs 150%, but EML need over 200% for that.

Just look at the next picture, and it becomes clear what happened. I spoke with this customer via Face time. He understood the manual in such a way that the setting had to be done with those knobs. Because actually, that is what is written there.

I asked him what to his opinion the meters were for. He said: Just to see if the tubes work.

Well the good part is, the EML tubes survived the attack.

VAC 30-30

Good amplifier, but it does put quite some stress on the 300B tubes.

We are in contact with the manufacturer to see if the 1.5 Ampere 300B-XLS tubes can be used here. Probably the answer is going to be yes. We are waiting for a confirmation. Please ask them yourself if you need to know.

VTL 185

Adjustment range is not very high, and (good) new KT88 tubes may draw not enough current. When replacing with Svetlana S-logo, Sovtek or EHX, it is recommended to use KT88 that show test values much above 60mA, by the factory test method. For other brand tubes we don't know the limits, but you need the higher current ones here as well.

VTL. Model: ST 150

Adjustment range is not very high, and (good) new 6550 tubes may draw not enough current. When ordering Tung-Sol 6550, it appears you need tubes that are tested above 60mA, by the factory methods. The same values apply for EHX 6550.

Welborne Apollo

This is a very nice amplifier, much recommended products! However the company has recently disappeared into a black hole, and there is no support now.

We see often, users want to replace the KR842 tube if used in there, by original 520B. Please take good care when doing so, because different tube numbers always indicate the tubes are indeed different. So by ignoring this you can have problems with the tubes or the amplifier.

Running EML520B-V3 in the Appollo.

You have to be careful with tube replacement, if you just replace the tube you find in the amplifier by the same type. Sometimes, mistakenly a wrong tube was in there. So that's why it damaged, and by simply replacing it, you repair the problem.

Sometimes the amplifier may have been modified to adapt for some other tube, and you would replace it perhaps by the original tube. Same problem.

Also there is bad and dangerous confusion with tube numbering of AVVT, VAIC, AND KR. For that reason we did not adapt this tube number system at EML. Though we build the same tubes, we gave them other numbers, so not to get involved in the confusion too.

25 years ago, this amplifier was made for the old VAIC VALVE 52B, but soon after it's introduction, the great Wizard Mr. Vaic was introducing variations in heater voltage and heater current. Yet one way or another, EML520B-V3 can be used, of the HEATER VOLTAGE IS CHECKED AND ADAPTED. This is regardless of what the previous owners have put in.

We recommend replacement by the EML520B-V3, but at replacement, two things have to be verified:

  1. The heater voltage. Do this with the old tubes first. It should be 5V, and adapt it down to 5V, if you find 6.3V or 6V inside. Please do not use the expensive EML520B-V3 as an experimental load, when changing the heater voltage! Better first use the old tubes for such experiments, and when that works, make a fine-adaption with the EML.
  2. The plate current must be 100mA. This is set by the value of R20. The current through R20 is calculated by ohm's law: Voltage over R20, divided by the value of R20. If you change the value, that will give a smaller change to the voltage and a larger change to the current.