12ax7 vs 12ax7a

Hi guys
I'm in the mood for tube rolling my Phono stage which uses 12ax7a valves.
I'm looking into going nos. Could I use 12ax7? What's the difference between 12ax7 and 12ax7a?

 

Nice one thank you

 

12AX7A is an "updated" lower noise version of 12AX7, and as Mr. Hoffman said it has to do with the way the heater is designed. Another 12AX7 "update" are 7025 tubes which are supposedly less noisy than 12AX7A and with lower hum, again because of heater being redesigned. With new production tubes al this means nothing. For example, TAD sales 7025 tubes but these are not actual 7025 tubes, they've just been SUPPOSEDLY selected from the current production 12AX7/ECC83 tubes for being less noisy but I seriously doubt that TAD or Shuguang that makes TAD tubes checks 12AX7 tubes one by one to get the quietest ones.
For NOS tubes, I have dead quiet 12AX7 tubes and some more noisy 12AX7A tubes,it depends on the maker,the raw materials used, their quality controls and very important, how these old NOS tubes were stored. The quietest 12AX7-family tube I have is a pair of NOS Hitachi 7025 that apart of being dead quiet up to the point that their noise level seems like solid state,their sound is really great, smooth with deep bass with real authority and controlled, warm mids and highs and not overly sweet, they sound nothing like solid state. I paid 35 $ for the pair includding shipping to Spain from the US.

 

I have a Pro-Ject Tube Box S and of course want to change the tubes.



Matched Pair PSVANE 12AX7 - T PREAMP TUBES . NOS , NIB Unfinished Project A are going for $75 (Australian) is it worth it? Seller has 7531 100% feeback.

 

Not true about "lower noise version." The 12ax7a was designed to be able to be used in circuits that used either parallel or series filament heater circuits. The original 12ax7 was only designed to be used in parallel filament circuits.

Since today you'll pretty much never run into a series filament circuit and pretty much only will see parallel filament circuits, the difference between a 12ax7 and a 12ax7a is functionally meaningless and, as noted above, in current production tubes, you can't count on their being any difference between a tube labelled 12ax7 and one labelled 12ax7a. But the "a" has nothing to do with a specification for "lower noise."

 

I have heard many a time that the "A" suffix means a slower more controlled warm up time. In fact the "A" variety 12A?7 tubes I have used take longer to glow and are slower getting to full brightness.
I have pretty much settled on Raytheon blackplates from late 50's-early 60's as my favorites. Some are "A" and some are not. They sound the same to me.

 

That's around 47 Euros, not expensive at all. I like 12AX7 and 6SN7 Psvane tubes so for this price I think they're worthy of getting this and give them a try. The ones I own but currently don't use on my Phono preamp,an EAR 834 "clone" which is a REAL tube Phono preamp, are Psvane 12AX7 MK II, I've never seen the ones pictured on the Ebay link but they look just like the ones I own on a different and not so looking box as the ones pictured.
I'm currently using NOS Matsushita 12AX7A or NOS Hit Ray (a 1960's joint venture between Hitachi and Raytheon) 12AX7A as I like tube rolling.
The Hit Ray 12AX7A are quite hard to find, I have two pairs and I'm always looking for more on Ebay, Matsushita 12AX7/12AX7A are much easier to find in NOS condition and they use to be inexpensive. I prefer the NOS Matsushita 12AX7/12AX7A than the Psvane but I could perfectly live with the Psvane 12AX7, I think they sound great and won't get rid of them as they will eventually be a good spare pair of tubes for my Phono preamp. I much prefer Psvane 12AX7 than new production Gold Lion 12AX7 with golden pins which I also own. I suspect the reason why I don't like how the sound on my Phono preamp, dull, lifeless and with rolled off highs, is a matter of sinergy between them as I also own a Yaqin MS23B Phono preamp and the Gold Lions sounded great with good detail and smooth and warm sound. Take your poison...
There's a "but", I suspect your Phono preamp is not an actual Phono preamp but a solid state one that uses 12AX7 tubes as "buffer tubes" most likely on cathode follower configuration to keep impedance low and reduce distorsion,and I'm not the only one here who thinks that about your Phono preamp. A simple way to check this out can be done by changing the 12AX7 tubes on your Phono preamp for a pair of 12AU7. 12AX7 tubes have an amp factor of 100,12AU7 have an amp factor of 25, if you put 12AU7 tubes on your Phono preamp and the output/sound level doesn't decrease, your Phono preamp is using the tubes as "buffer" and not for amping the signal of your cartridge.

 

Correct. People with the same phono stage have often spoken about getting a better sound when upgrading those very basic tubes. I do not expect to be blown away but do think I can make a fairly cheap upgrade worth while.

 

Why don't you try a pair of inexpensive (even better if you can return them) 12AU7's instead of the 12AX7? If used as "buffer tubes" on cathode follower configuration, tubes that keep impedance low and have more current to spare if necessary are better suited for this task and I think 12AU7's are better for this than 12AX7, and they don't mess up your Pro Ject Phono preamp. The EAR 834 Phono preamp, which I own a clone, an exact copy of its circuit design and specs, uses, according to its designer, three 12AX7 tubes. V1 tube does around 75% of amplification for BOTH CHANNELS, V2 tube does RIAA de-eq and the remaining amplification, and the V3 tube is used as a "buffer tube" on a cathode follower configuration to do exactely what tubes do on your Phono preamp, keep impedance low to reduce distorsion. I'm not using a 12AX7 tube as a V3 tube, as specified by the maker, but a 12AU7,it just does a better job at keeping impedance low and has more current to spare ('though it has less gain than a 12AX7 but "buffer tubes" are not about gain), the EAR 834 "clone" sounds better with a 12AU7, it doesn't damage the Phono preamp and you can get NOS 12AU7 for dirty little money,they use to be cheaper than 12AX7. I'd give 12AU7's a try.

 

12AU7 Preamp Electron Vacuum Tube 9-pin Dual Triode for ECC82 5963 5814 E5B5 | eBay

Will this do?

 

Yes, this will do but I'd get a pair of better quality. Let me do a search on Ebay for you.

 

I'd get a pair of these,they are cheap, only 8 GBP per piece,they're NOS and military grade tubes which means they must be well made.
5963 Philips ECG 12au7 Ecc82 Tube Valve Old Stock | eBay

 

That looks quite good, will get it.

Thank you so much for your help!

 

I'm here to learn about the differences, but do not understand how a different filament type could make any difference in a series circuit?

 

It's a good question. I'm repeating what I was always told for years as a purchase, but as someone who spent 25 years in the news business, I ought to know better than that. I'm going to go back to the datasheets to see if there's actually any difference in how the heaters are wired.

 

You're welcome. Will you tell us your opinion about the sound of these when you get them?

 

Even if it did make a difference, all quality tube audio gear has the filaments wired in parallel. As far as I know, the series wiring was for economy only, found in radios and portable record players which have no power transformer.. so that the filaments are powered directly from the 120v AC supply. The series wiring drops to the correct voltage for each filament in series, and shared in series to some other electrical component such as an AC motor, pilot lamp, or resistor.

 

One 12AX7-kind of tube that had filaments wired in series are 7058's as they were intended for use on portable units like car radios. I used a GE pair of them on a Yaqin MS23B Phono preamp and they sounded great.

 

When I read about these matters (low noise devices) etc, we EEs know what it really means, but isn't the same as what someone who doesn't deal with these matters professionally. Just wanted to explain what noise values really mean WRT vacuum tubes (and transistors with different scale factors/etc.)

There are a couple of audio-important noise sources when using electronic amplifying devices, one is the natural (thermal) noise of an equivalent resistance, another are excess noise sources other than 1/f, and then 1/f noise sources. Often, the noise sources are simply put together as a single source on old tube manuals and called 'equvalent noise & hum'. The intrinsic (natural) noise source is something that just cannot be worked-around, and that is the noise of an equivalent resistance (it is specified differently for bipolar & FET transistors, but is essentially the same.) If one isn't working directly in the audio band (not passing the hum frequencies), then using the thermal noise numbers alone can be good guidance.

If one purchases a good 'tube' like a 7025, they state an equivalent input noise as something like 1.8uV typical, and 7uV max over 10kHz. However, that noise is made up of the naturally necessary thermal noise of the tube, and also extra noise sources.

Non-EEs might not realize that the naturally necessary (my term coined) thermal noise of a tube or transistor is usually inversely related to the gm (transconductance) of the device. The higher the gm, the less the minimum possible noise of the device given everything perfect. Transistors & FETS nowadays tend to approach those minimums because of excellent fabrication processes, but tubes (valves) are generally stuck with the technology of the past, and also have lots of heat (and hum) sources that naturally generate extra noise. However, the minimum thermal noise related to the gm for a triode is approx 2.5/gm, which on a 12ax7a/7025 type part gm approx 1250umohs, or the Req calculates to be 2000ohms. This effectively means that the input of the tube has the noise as if there was a 2000ohm resistor in series with the input. That equivalent noise is approx 0.7uv for a 10kHz BW.
What this means, that even given an ideal tube with NO extra noise sources will have an effective input noise of approx 0.7uv over the band of approx 10kHz. (Req for transistor designs can be anywhere from less than 50ohms up to 10ks of ohms... As a first order optimization, using higher gm parts or using parts in circuits that manifest higher gms can be helpful, but on transistors there is also a kind of parasitic resistance, rbb, which puts a limit to the increase in gm.)

The 'real world' spec of 1.8uV or the max of 7uV adds in the various other noise sources that the real world causes. The 'perfect' audiophile tube could theoretically approach 0.7uv assuming everything else was perfect, and the limitations of the heating method would be ignored. NO TUBE WILL EVER DO THAT. However, transistor designs (for audio) can often come close to the theoretical minimums. It is a challenge to approach noise limtiations with tube designs, and such design is a real art & science mixed together. The relatively high Req of most 'tubes' imply that they work with lower noise figures at higher impedances.

I don't know the LF characteristics of the specific tubes/valves that are being talked about, but usually the largest noise density (not necessary noise total amount) is in the lower frequencies, where then the noise density rolls off at higher freqs to a constant up to certain -- usually significantly higher frequency -- then the noise goes up in value again. This is interesting (and requires some evaluation) for typical MM phono RIAA designs, because the gain is usually much higher at the lower frequencies -- where the amplifying devices tend to be the most challenged (including dealing with hum.) Believe it or not, assuming a reasonable design, high audio frequencies for preamps aren't as much of an issue. LF for RIAA tends to be slightly more challenging. The saving grace about the LF range is that it is much narrower also, so less room for the noise to be created.

Basically, I wasn't really trying to teach everyone about low noise design, but to make it more obvious to non-EEs that there are REAL minimum limits that most electronic amplifying devices must produce. Devices like tubes, transistors and FETs have similar kinds of limits. (There are devices called 'parameteric' amplifiers and other exceptional cases which also have limits, but they are a different kind of limit.)

John

 

Of course. Will get the car paid off this year and from next year will put money aside for a proper tube amp and preamp. To think of all the dumb things I spent thousands on when I was younger just makes me mad now.

 

Also -- forgot to mention WRT some other devices (like BJT -- bipolar transistors), which also have a substantial input bias current, which also causes a substantial noise current component -- so that unlike using 'normal' impedances at 'audio' frequencies with FETS and 'tubes' which have primarily a voltage component to the input noise power, BJTs also have a current component. SO, with BJTs there is an ideal source impedance for noise, while with tubes and FETS, the 'ideal' source impedance is sometimes very high because the effective bias currents are so very low.

To avoid confusion, the ideal input noise for a transistor/FET or tube is actually a noise power. With tubes and FETS at medium frequencies, the idea of ideal source resistance can mostly be ignored in audio applications. Maybe there are cases of condenser microphones where the impedance might be high enough where the noise current of a FET or tube might be high enough with the source resistance to create more of the in*R noise term. So more completely -- an input device effective noise voltage is vntotal = en*sqrt(BWHZ) + in*sqrt(BWHZ) * Rsource.
Interestingly -- to me, one reason why devices become more noisy at higher frequencies is that the 'in' term increases as the frequency increases. It isn't necessarily that the increase in noise at higher frequencies is just base upon 'en' alone. The same goes for frequencies below the 1/f en and 1/f in knees. (that is, for lower frequencies the noise terms increase also -- but not as bad as it sounds because the actual bandwidths are narrower at lower freqs.) The bad-old 'popcorn' noise which used to be common when using older style transistor processes was even worse being of 1/(f^2) character or worse.
All amplifying devices (tubes transistors) have all of these terms (except for some kinds of amplifier configs like chopper amps/parametric amps/etc), even tubes having the in (input current noise) terms, it is just that they aren't as important under most common audio situations. AT RF, yes -- the in terms increase.

John