Brian Fox 11/24/15
You would think that enough has been written about the u47 to pretty much put the topic to bed but no. I had the opportunity to do a re-build on good clone and I had to report what I found if only to keep it from echoing around in my head for all eternity. I also needed a way to give some creds to Telefunken USA who lent me a brand new VF14K tube. It was invaluable to the process. Thanks!
Once Upon a Time
I received a u47 clone that was built by the late Oliver Archut, an expert in the field of transformers and classic microphones. It was a very good replica of the original including all the things one would expect. It had an m7 capsule built by Gefell, a BV8 transformer that looked and measured like an original, the circuit was accurate with a real VF14 tube and it used a single power supply to feed both the B+ voltage and the filament. It also boasted a proper u47 detachable head-basket with the built-in cardiod/omni switch. It was as close as we can get to building a u47 in the 21st century.
The owner of the microphone had lent it out and it came back looking good but after running for a short time it made noise. The story was that someone forgot to turn off the power supply before plugging in the microphone and the mic got a blast of abnormally high voltage. I did not know it until later but this power supply was bomb waiting to explode on this microphone.
The owner, who is mending after some medical problems is on a tight budget and could not afford to replace the VF14 and was also concerned that the money spent could be easily lost if this happened again. The microphone was sent to me with some NOS EF802 tubes and requested that I modify it for the new tubes. The client had been told, as had I, by Oliver, that the EF802 had the closest characteristics to VF14. I made a point of using this opportunity to try and discover the accuracy of that belief.
When everything arrived I connected it up to listen to the noise problem to see if maybe, just maybe, it was not the VF14 tube. Given the price of an old VF14, it would be worth spending the time to rule out the tube, before going ahead with the rebuild for the EF802.
After connecting everything and listening to the mic, everything sounded clean. I did not hear the classic bottom end I expect from a u47 and made a note of it for later. Sure enough after fifteen minutes a little noise started to creep into the sound and within five more minutes it was continuous.
My impression was that there might be oxide on the ancient tube socket so I removed the tube and cleaned all the copper contacts with some gentle pressure to be sure they were clean. No change.
I checked all the solder joints and reheated them from top to bottom on the off chance that one of them was a cold joint. No change. The noise came back every time.
I checked the headbasket assembly, the cardiod/omni switch and the associated joints. Nothing.
I even adjusted filament voltage up and down to see it that would help. Nothing!
Maybe the power supply had the noise problem. I scoped the supply and could see nothing. It was a choke filtered, four stage filtered supply. It was beautifully clean.
The u47 Circuit
With the microphone opened I was able to measure voltages around the circuit to make sure they were within the norms for a u47. The filament voltage was definitely much too low which was strange and it also meant that all the other measurements were irrelevant until I got it correct.
The u47, at first glance might look like it was designed to save a little money. In an era when parts cost a lot, if you find a way to reduce parts like an extra power supply, you added something big to the profit on each microphone.
The u47 designers did something that was more like what I saw in old table top radios and tube televisions from the 1950s. In the old days home radios and TVs were designed using tubes that had higher voltage filaments (the glowing part inside). This was done so you could wire the tubes in series and connect the filaments directly to the 120V AC power from the outlet and save the cost of a filament voltage transformer.
Professional tube equipment on the other hand, in my experience, always used a separate low voltage, high current supply for these tube heaters although it was still AC voltage.
The side-effect of that AC filament voltage is that a small amount of AC hum can get into the audio because the filament is close to the cathode electrode inside the tube.
In the early 20th century it was near impossible to make a high current, filtered DC power supply so the innovation of the u47 was to simply use the nice clean high voltage supply (called B+) to do double duty for the amplifier and the filament.
The u47 brings in 110V DC to the mic to feed the amplifier circuit and uses a big resistor to drop the voltage going to the filament. This is only possible because the VF14 has a high voltage filament, nominally 50 volts, which means it needs much less current.
So even though the u47 sounded fantastic, a testament to the designers, there is a number of side-effects of running the whole mic on one power supply. One is that the tube filament needs 90 times more current than the amplifier needs. This means the 110V power supply needs to be pretty beefy and it requires extra filtering since it has to produce all that current with little to no hum in the power.
Another side-effect is that every part of the circuit affects every other part of the circuit. Since the supply is not regulated if you draw more current the voltage goes down, which affects the filament voltage. And since the filament is used to create the bias voltage that effects the amplifier operating point, which affects the head room. Oh and by the way, it will reduce the capsule voltage too so the sensitivity goes down. It all has to be correct for this microphone to operate correctly. It’s like a very crazy feedback loop. Funny thing is, nobody would design a circuit like this today, yet as we know a properly tuned u47 sounds amazing.
The Ticking BOMB
My first measurements were made with the u47 plugged in to the supply. Everything looked pretty normal. The filament voltage was a little lower than the Neumann drawings but it was ok. While inspecting the supply I disconnected the mic. for the first time and measured some voltages and what I found pointed to what I believe was the root cause of the noise problem. A tube power supply is a beast and they need a little taming. A large set of filter capacitors in a power supply it is capable of holding a lot of electric charge for a very long time.
These large capacitors are a good thing to filter the AC ripple out of the power. They are a bad thing if you don’t provide a controlled way to discharge these capacitors when the supply is turned off. The simplest way to do that is with a thing called a bleeder resistor. The bleeder resistor lets the capacitors discharge to ground in a few seconds so that when you switch off the supply, it does not hold all that voltage for days. (which it can do without a bleeder)
The other thing the bleeder does in a non-regulated supply is sets a maximum output voltage that the supply will generate with nothing connected to it. Do you see where this is going…?
After unplugging the power supply and waiting for about a minute, I started poking around the wiring to better understand how it was wired.
When I regained consciousness…
Well not quite. It was only a marginally less-than-lethal shock but now I was sure that there was no bleeder resistor. It made me wonder, how high was the voltage with the microphone un-plugged?
The answer: 402 Volts! Yikes!
So although I cannot prove it I am pretty sure that was the cause of the mic failure. Even unplugged from the wall socket, this power supply would still have a lot of power stored in the filters and it was a very high voltage. If anyone will give me another vintage VF14 tube I can do some tests to prove it 100% :-)
If a microphone in the u67 style interests you send us a note to request a quote at: