PX4 -or the big confusion with the tube data

In search for the PX4 tube datasheet, I was surprised, it was missing the essential numbers for gain, transconductance and plate impedance. At a first sight, they seem to be in the datasheet, but this is at the "Emission" test point, with zero grid voltage, and as we all know very well, gain, transconductance and plate impedance depend heavily on the operating point. Though this can be extracted from the tube curves, which are so to say the DNA of the tube. Yet, the original old datasheets does not specify what is the transconductance, plate current or gain of the PX4 at any normal working point is supposed to be. All that is available is a curve chart from OSRAM, which is detailed and precise. I was never able to find an explanation why this data has been missing ever since, and why nobody ever tried to restore it. Even though the Chinese build what they say is a PX4, they either also do not give this data, or rather data which is plain wrong. The intention of this article is, to come up with the required data for the PX4, and also have a critical look at what companies like KR Audio and the Chinese have been doing, while building tubes and giving it the PX4 part number.

Call for information. If you have information about the PX that you would like to place here, just send an email, we will be glad to place it here. Also, there could be errors on this page. If you think so, please let me know as well.

The socket. The PX4 is a tube with the British-4 socket, also called B4. This is actually a good socket. I was never able to fin an accurate drawing for it. Yet all tubes seem to have nicely the same dimensions. The difficulty with the B4 socket is still bad pin contact with very old tubes or sockets. Reason is, the pins itself had a (very nice actually...) spring inside each pin. The intention of that was to have better contact. However, the wire coming from the tube, usually is soldered at the tip of the pin, and so the wire is lead along that spring contact. This means the spring contact is contaminated with solder flux, and after enough decades that does seem to corrode the pins, and by that the tube sockets too. So in some really rusty old radios, the sockets are usually totally out of function. The pins of most tubes are heavily corroded, which can be cleaned, but you can't clean the old sockets So one way or another, though I think the socket is nice, it was never very popular. Some very nice Telefunken tubes have the wires soldered directly at the tube base, so to prevent solder flux damage with the tube sockets. These are rare however.

datasheets. What I find I will put in the next list. The older datasheets are from Marconi. The later datasheets are from Osram, they look so much the same, they are rather identical. So that means there is only one data source, be it Marconi or Osram.

Missing data. From looking at the datasheets, you will see that transconductance, gain, and plate impedance are specified at 100 Volts Anode, and zero Volts grid. Now this is really strange, and a bit wrong, because at zero volt grid voltage the tube cannot amplify AC signals, so what's the use of specifying it that way? I just don't know!

Detail from the left page

As you can see in the above picture, this datasheet if for the earlier 12 Watt tube. Later, datasheets were 15 Watt. The 15 Watt tubes can replace the 12 Watt tube. The 12 Watt tubes can only replace a 15 Watt type, if the actual amplifier does not bias the tube above 12 Watt. So this needs to be looked at with great care by the user when using vintage tubes.

This is already the 15 Watt tube.

From this datasheet, it sais:

Measured at Va = 100V, and VG =0V:
Mutual conductance = 6mA /V
Gain = 5
Plate impedance = 830 Ohms

Though of course, nobody will operate the tube at 100Volt anode, and zero Volt grid, as that would bias the tube at 65mA, meaning 6.5 Watt dissipation, and also you cannot put an AC voltage on he grid anymore when the grid at zero volt. The reason for these "funny" values are another. This working point is a so called "Emission Test", and it is a perfect test for power tubes. For more details, read about the Funke W19 tube tester. This method tests the tube in diode mode, doing so at relatively low voltage. A weak cathodes may still work reasonable at higher plate voltage of 250...300V, at only 48mA, but have a problem at lower voltage. This problem gets even larger when the required current is much higher. The PX4 is an English tube, and the famous AVO tube testers, worked with a real high power Emission test. So this setting of 100Volt (at Ug=0V) and requiring 67mA like that, you find it directly in the AVO Mk1 tube tester settings, and also in the PIOTR bible. By changing the grid voltage from 0 to -1Volt, the transconductance at this unuseual working point could be obtained, which is a very high 6mA/V, as the 0Volt curve is the steepest of all curves of a chart.

Yet this was represented in the datasheet a bit unclear at a first sight, and many are confused, thinking PX4 is a 6mA/V tube. Well it is, but at zero Volt grid, and at 100V / 67mA, and you cannot use it in an amplifier that way. At 250V /48mA it comes down to rather 5mA/V. The Chinese specify for their recently made PX4 6mA/V but are doing so at normal working voltage. This shows how quickly one can be mistaken. It can be seen, they used all data of the 100Volts emission test, and mistakenly specify that for the normal 250...300V working point.

Sometimes my PIOTR bible will help, but here it also doesn't help much. Though it does give some kind of a clue. The famous test point at Ua=100V is declared here "STAT". Though I do not know for SURE what the means, if it means "static data" or so, that is unlogical, as they write dynamic parameters there, like transconductance. Yet we can see here, they declared this test point not as as class A or Class B test point, but as a STAT test point.

This is a list of tubes called to be equivalent with PX4, but I do not see it like that. All tubes have similarity, but at some points, the parameters are so wrong, I would not advise those as a serious replacement for PX4.

	Ug	Ua	Ia	Gain	Rp	Gm
PX4 Marconi Amplifier use	-32V	250V	48mA	5.6	1133	5
PX4 Marconi Emission Test	0V	100V	72mA	5	830	6
AC044 Emission Test	0V	100V	52mA	4	1150	3.5
LP4
P12/250 (data unconfirmed)	-27V	250V	49mA
PX41
RE604K	-23V	250V	48mA		830	4
S30C
SSHV42
4XP

Original Osram Specifications in brown Missing data reconstructed by myself, in green, from above OSRAM curves. So far I was not able to find official (historical) information for that. Please let me know if you find something.
Anode Voltage	300V (Max)	250V	200V	100V
Grid Voltage	-42V	-32V	-28V	0V
Anode Current	50mA	48mA	25mA	65mA
Anode Dissipation	15Watt	12Watt	5Watt	6.5Watt
Bias Resistor (AC heated)	900Ohms	740Ohms	1200 Ohms
Load Resistor (Ra)	4000 Ohms	2400 Ohms	4500 Ohms
Gain (Mu)	4.9	5.6	4.6	5
Transconductance (Gm)	4mA/V	5mA/V	3.2mA/V	6 mA/V
Plate Impedance (Rp)	1233 Ohms	1133 Ohms	1448 Ohms	830 Ohms
Estimated power Output	4.2 Watt	3,8 Watt	3.5 Watt

This chart and electrical data was found on the Sofia Electric Website on October 2016

Sophia Curve Chart. Please don't be confused, "SOFIA" is a model name of very beautiful Italian tube tester, which was used to extract these curves from the tubes. Whereas "SOPHIA" (with the P) is not a tube factory, but a trade mark in the USA, selling Sofia branded tubes made by Full Music in Tjanjin, China.

They do not have tube specifications on their website, but from the test results of the Sofia tester, I quote this data:

Measured at Va = 300V, Ia = 52mA:
Mutual conductance = 3.27mA /V
Gain = 5.4
Plate impedance = 1.540 Ohms
Grid Voltage reads -45V from the curves.

This chart and electrical data was found on the Fullmusic Website on October 2016

Fullmusic Curve Chart + datasheet.

They have a tube chart on their website, as well as published data. Note, the SOFIA tester, as used also extracts the dynamic data from the tube as well. So that is two sets of data, Fullmusic presents. I noted, these two sets of data are quite different.

Published as text:
Measured at Va = 300V, Ia = 50mA
Mutual conductance = 6mA /V
Gain = 5
Plate impedance = 830 Ohms

The above is quite different what their SOFIA test results say:
Measured at Va = 300V, Ia = 52mA:
Mutual conductance = 3.27mA /V
Gain = 5.4
Plate impedance = 1.540 Ohms
Grid Voltage reads -45V from the curves.

They specifically write under the Sofia curves, these are made with AC filament voltage. I do not beleive this, I have two Sofia testers myelf, and the schematics of them. I can say for sure, the Sofia tester has only DC heater voltage, and no option for AC voltage.

Extracting transconductance from the chart.

For extracting the data, as I did here, I had to make one curve a bit longer. (It is the green line) I found a value if 44mA / 18 Volt = 2.44mA / V that way. However, the Sofia software will not do so, and instead of that takes teh chart "as is" and this introduces an error, while trying to construct outside excisting data. You can easily see this, when you slide the cursor along a curve, and when the Gm reading in the right corner makes suddenly a big jump, you are too close to the edge. It seems happended here, as the marker is quite high up the -45 Line. This is to my opinion why the Sofia comes up with 3,27 mA /V, while I see only 2.44mA / V when I verify it from the chart. I consider the 2.44mA / V correct.

Conclusion: Transconductance of the Chinese tubes is much lower as they write. I measure 2.44mA / V from their website chart, wheras they specify 6mA /V.

KR Audio

Published as text, as seen on their website, October 2016:

No curve chart.

Tube data:
Measured at Va =250V, Ia = 48mA
Mutual conductance = 5.1mA /V
Gain = 5
Plate impedance = 980 Ohms (See Note)

Note: Plate impedance = not specified by KR. However, by Barkhausen Formula, plate impedance = Gain / Transconductance. So I calculate 980 Ohms in the above working point.

Conclusion: Other than the Chinese tubes, the KR published data is nicely corresponding to the original OSRAM tube curves.