• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). There are daily reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Carver Crimson 275 Review (Tube Amp)

Rate this amplifier

  • 1. Poor (headless panther)

    Votes: 356 95.7%
  • 2. Not terrible (postman panther)

    Votes: 5 1.3%
  • 3. Fine (happy panther)

    Votes: 4 1.1%
  • 4. Great (golfing panther)

    Votes: 7 1.9%

  • Total voters
    372

captainbeefheart

Active Member
Joined
Apr 18, 2022
Messages
276
Likes
356
Kevin Deal seems like a decent enough guy, I'd need to hang with the cat in person to get a real vibe from him but his videos are loosely entertaining, informative maybe. He talks about the "resistance" in the filament and that's why it flashes but that's a very broad brush stroke for an answer. It's really to do with the fact there is a piece of the filament that's not coated and sticking out of the cathode and although it has to do with resistance, it's actually the lack of resistance in a cold filament, on startup it pull a large surge current because of the cold low resistance of the filament, the flash is from the part of the filament that's isn't coated and sticking out from the cathode. Regardless he's for the most part correct it's not an issue with performance or life.

The simplest way for me to explain tube life to a laymen is focused around the cathode, which is where the magic happens. The filament, which is also the cathode is heated via the heater winding which is in close proximity inside it, this is why there is a coating on the filaments in order to mitigate shorts between the two. Once the cathode is heated to a specific temperature the metal starts to change it's behavior. We all learned in school about the outer valence electrons and why metals are conductive, these electrons begin to be "boiled off" and create a negatively charged electron "cloud", this is known as emissions. The cathode emits electrons. These negatively charged dudes are attracted to the highly positively charged anode called the plate, the control grid between the two can control this current flow between cathode and anode. Back to the "electron cloud", you reach a point where when the cathode emissions get reduced, with less emissions you get less current flow hence a lower transconductance reading; I.e. Given the same change in voltage at the control grid you will have reduced change in current from the cathode to anode. This is where I feel the vast majority of tube vendors get it wrong and you the consumer pay for their error. They believe the reduction in emissions happens in some linear fashion in relation to time, I'm sure you have heard it before, this tube is 90%, Kevin in the video most likely screwed over his customer by selling him those KT150's because in his mind the tube still has 90% of it's life left, that's not how it works. It's transconductance is 90% of average value for the given tube type, that's all that means and nothing to do with life, in fact tubes stay fairly consistent in regard to their emissions and subsequent transconductance for their majority of their life right up until the end, then once emissions starts to become reduced the tube doesn't have much time left until the cathode is completely depleted and junk. The fact Kevin measured a reduction in transconductance from it's original value means that the tubes are most likely getting tired and on their way out, it doesn't mean it has 90% of it's life left, it doesn't linearly drop in percentage the more you use it.

So how can one tell if a tube is getting tired? There can be a +/- tolerance in transconductance on brand new tubes so how can one tell if there is a reduction in emissions that's not so obvious. When I say obvious is say you have an EL34 and it measures 4000uMhos one can assume it's going in the trash can, but if it has say 9500uMhos is it a new tube just lower on the end of gm for the type or was it originally 11000uMhos and now down to 9500uMhos and so showing a reduction? How does the cathode work? It needs to be hot right? What if we play with the temperature of the cathode to see how weak it is, can we do that? Sure we can, well sort of. A strong tube will still function well with a slightly reduced amount of heater power, notice I said power. My tester uses a current source for tubes, so I reduce the amount of current through the heater while watching transconductance, if gm drops with slightly reduced heater power then I feel confident saying the tube is in the later stages of it's life because a strong tube won't budge.

I don't know Kevin, he seems like he means well and wants to help people but he has made quite a bit of money helping people and I have learned I often don't profit much from "helping". I don't knock people who make a living and I don't think he is a malicious person, I'd actually love to hang with him and listen to music and chat about this stuff because I think he genuinely wants to learn knowledge. I really think tube vendors could do much better at what they do but I don't expect them to be engineers, they are sales people and that's it. I am just not a fan of ebay and the vast majority of vendors using transconductance as some sort of life scale % when it's far from that. Tube life is complicated yet not, my method of reduced heater power and a drop off of emissons has been the de facto method of finding weak tubes for a very long time, yet I never hear any of these upscale tube vendors mention a reduced heater power test for their NOS sales, especially used tubes.


Kevin doesn't mention anything about heater power which has a huge part to do with tube life. Tubes last much longer when run a little on the low side for power, but burn out much faster when run with too much power. This is a much larger factor vs if it's running at max dissipation, his Genelex example only went up to 10,000 hours for 66% dissipation vs 8,000 hours for 100% dissipation. As long as you are not going beyond the max limits of the tube they last very long and Class A vs Class B isn't as huge a deal as people make it out to be so long as you don't go beyond the max dissipation. Now heater power is the elephant in the room, keep preamp tubes at 6v instead of 6.3v and watch how long they last.
 

captainbeefheart

Active Member
Joined
Apr 18, 2022
Messages
276
Likes
356
In my homebrew design, I use a soft start circuit on the heaters to soften the inrush blow.

This is my experience with death of tubes, surge at power on finally stresses the weakest point in the filament and burns it open.

When I design for DHT amps I take great care into the filament power, I finally switched over to a current source because I found most DHT deaths were open filament, especially modern DHT's. When I switched over to a current source and even retrofitted some earlier models for local customers the feedback is starting to come in that the failure rate has been significantly reduced. Of course I needed to wait years to finally confirm the suspicion but I can safely say now the current source design amps have had zero filament failures to date while I see more and more amps come in with dead 300b's and an open filament, these are always AC/DC voltage source fed supplies. It's also no surprise failures are common when I see the supply voltage over 5v, typically 5.2 and as high as 5.5v, that's going to stress the filament. I set the current source for 1.1 amps and no more is needed, datasheet says 1.2 amps. At first I would adjust the current supply to reach a perfect 5v which works well but I found that so many were getting to 5v with less than 1.2 amps that running them 1.1 amps almost always kept the supply just about perfect and very long life without any reduction in performance.

I don't take lightly that someone may purchase some Western Electric 300b's and I won't sleep at night knowing my designs may cause these expensive tubes to fail when it could have been avoided. Running slightly reduced current via a current source completely eliminates surge current and allows better regulation of power to run them slightly reduced vs slightly increased, the latter just eats tubes up while the former extends life at no cost to performance.
 

egellings

Major Contributor
Joined
Feb 6, 2020
Messages
1,798
Likes
1,319
The only problem with using the I-src for a filament supply is that the heater current for the tube is not as tightly spec'ed as the voltage is, and you could end up running the heater at too low or too high a filament voltage, which is a spec'ed parameter for the tube. What I do is have my heater supply start out in constant current mode, and when the voltage reaches the specified heater voltage, the supply transitions to constant voltage mode at the specified heater voltage. That completely eliminate the inrush current problem, which is the real tube life shortener.
 

captainbeefheart

Active Member
Joined
Apr 18, 2022
Messages
276
Likes
356
The only problem with using the I-src for a filament supply is that the heater current for the tube is not as tightly spec'ed as the voltage is, and you could end up running the heater at too low or too high a filament voltage, which is a spec'ed parameter for the tube. What I do is have my heater supply start out in constant current mode, and when the voltage reaches the specified heater voltage, the supply transitions to constant voltage mode at the specified heater voltage. That completely eliminate the inrush current problem, which is the real tube life shortener.

Yes exactly, that's how I was doing it, bring current up until it reaches stated voltage. The way it was explained to me from an old Philips tube engineer was that current is more of an accurate method of optimal filament power. For example, a 12AX7 with heaters in series is 150mA, he explained that you're better off with running 150mA through it regardless of the voltage. Going by voltage only you could set it for 12.6v yet due to a slightly lower heater resistance you can easily be running 170mA. He said whatever the voltage is at the specified current of 150mA is what the tube needs. So he stressed making sure the opposite of what you're saying which makes sense. Him and many others said that a little under the specified current is actually fine for most tubes and will make them last longer without effecting performance. If you are at 150mA and only measure 12.3v across the filaments he didn't recommend increasing the current until you reached the 12.6v, he said specifically wherever the voltage is at for 150mA was optimal.
 
Top Bottom