• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

Help with Odd 5670 Tube Testing Results (B&K 747)(6922 Adapter)

Gestetner

New Member
Hi all, it's been forever since I've posted here. Lots of life stuff since then. Sorry for the long post.

I'm having difficulty testing GE JAN 5670W tubes I ordered from a reputable tube vendor, along with the 6922 adapter they sell. I bought 38 tubes and 18 adapters so that I'd have two full sets to populate the 6922 locations in my Sonic Frontiers Line 3 preamp (10x) and two Power 3 monoblock amps (4x each=8). The vendor is willing to replace the tubes I'm having issues with. I do, though, want to get to the bottom of why I'm getting the results to make sure the tubes are dodgy and that it's not just a weird quirk with testing this particular tube model, etc. Or an issue with my tester. I'd also like know of any tips to reliably test them if there's something else I can try.

I received the tubes just about two weeks ago. I went to test them on Sunday and I got a lot of troubling test results. Odd behavior. My tube tester is a B&K 747. It was professionally restored and calibrated by Doutt Electronics in 2002 ("Cap & Semi & Res Kit" + new sensitivity pot).

Note: Most of the testing I did was with the tubes as 5670's without the adapter. That is because they all test Bad when tested as 6922's using the adapter (testing from 25 to 39 in the Bad range). I've asked the vendor about why that is but have not received an answer back yet. (I do see that the 5670 sensitivity test setting on my B&K is 60 whereas it's only 24 for 6922 tubes. It might be that that alone causes all of the 5670 tubes to test bad as 6922's using the adapter. I'm waiting to hear the vendor's explanation on that. I would hate to think they would be "weak" in a 6922 application but don't know enough about tube/testing "physics" and what the test results mean to know one way or the other.)

Some tubes test fine. Allow a little warm up and the needle moves steadily to the Good range and its value. This is the type of test result that I've always gotten with this tester. I took some of my known good 6922 tubes and tested them to compare and they all followed this pattern (all tested Good, too). Some of the 5670 tubes tested consistently Bad (really low in at least one section) and those are definite rejects. However, about half of the 5670 tubes are giving me Bad or ? results, but also odd and inconsistent results. To make it easier, I'll separate points into numbered headings.

1. In contrast to normal test behavior (of moving steadily to their value), a lot of the tubes have a distinctively different test result pattern. Upon pressing the test button, the needle will quickly "Pop" with a jolt to a higher value then immediately drop to a worse value, often in the ? or Bad range. For instance Tube 35 jolts to 65 (? range) the immediately drops and sags to 47 (Bad). Tube 7 Pops to 65 then Drops to 41 or 42.

2. After getting a lot of weird results like this, I retested a few of the earlier ones that tested normally and they continued to test normally.

3. On tubes that Pop and Drop as described in #1 above, I will sometimes get worsening performance over immediate retests. (Varied wait times, maybe 5s, maybe 10s, maybe 30s.) For instance, Tube 24 1st Section: Pop to 80 Drop to 70, repeat test and Drop to 75. Retested the next day, the results were Pop 70 Drop to 59, then subsequent tests settled at drop values of 61, 63, and 65.

The next tube I tested (Tube 25) tested in the normal manner by steadily rising to 97 on the first section and 98 on the second. No Pop, no Drop.

4. On many tubes that Pop and Drop, if I hold the test button in for a long time the value very slowly rises, dramatically. By "long time", I mean like 20-30 seconds or more. For instance, Tube 30 Popped high, Dropped to 70 but with a very long press of the Test button (30s or more) the needle rose to 85 where it topped out and sat until I released the button.

Tube 35 was sometimes similar but also odd. 1st Section: Pop to 80, Drop to 65, extra long hold and it slowly rose to 82. I waited 30s and retested without a super long hold and it settled at 70. 2nd Section: Pop into the 90s, Drop to 82, Hold rise to 97. Wait 1 minute, retest, Pop to 92, Drop to 83, long Hold to 95. However, I then retested it right away without long holds and it settled at 60, then 60 again, then 60 again. What the heck?

5. I have only gotten "Good" test results on slightly less than half of the tube order. That is surprising because I thought NOS JAN military spec tubes from 1984 should be decent quality. The vendor said they were unaware of big problems with these tubes but who knows if people test them right away.

I can just return the ones that are giving me weird test results and the vendor will replace them. However, I don't want to be caught in a never ending loop of testing weirdness on the next batch. Getting odd results and retesting over and over again is very time consuming. So I'm trying to see if anyone here has any insight one way or the other on what might be going on or other steps I could take to figure this out. Like I said above, my B&K 747 has been working fine and gives me consistent results other than these tubes. I've used it to test 12AX7s, 6550C's, 6L6s, 6SN7s, 5U4GBs, 6922s and more without any problems. I would also like to be able to get a consistent testing regime so that I can sort the tubes for use as pairs in the sections of my preamp.

Thanks.

p.s. I am in Orange County, CA if anyone else lives nearby and has a reliable tube tester that I could cross-check some tubes on. I looked at the sticky for other OC people and didn't see anyone current. One was from 2005 and the other hasn't visited since 2012. I even searched the sticky for "California" and the remainder were in NorCal hundreds of miles away.
 
Register to hide this ad
A 5670 has less than half the transconductance of a 6922. I don't see how they could ever test the same. Was this an eBay sale? If so, I would send the entire lot back for refund, tubes and adapters. If it was a direct retail sale, tell the seller many of the tubes test bad, and ask if he will test/replace the bad ones for you, or if he prefers to just take them all back (including the adapters).

Jack
 
A 5670 has less than half the transconductance of a 6922. I don't see how they could ever test the same. Was this an eBay sale? If so, I would send the entire lot back for refund, tubes and adapters. If it was a direct retail sale, tell the seller many of the tubes test bad, and ask if he will test/replace the bad ones for you, or if he prefers to just take them all back (including the adapters).

Jack
Hi Jack, no it wasn't an ebay sale. I bought them from one of the big tube suppliers, TubeDepot.com. This is the product page (I got the combos and tubes on sale so I didn't pay this price):


I spoke to them right away about the fact that they test Bad as 6922s and have been waiting over a week for an answer. I just contacted them yesterday about this weekend's really poor results testing them as solo 5670 tubes. They offered replacements and a return shipping label for whatever I wanted to send back. I'm trying to make sense of the testing, and using these tubes, before deciding what to send back. And, how to test the remainder and any replacements. I definitely raised an issue with testing within the two weeks but I'm not sure if they would accept a full return at this point (I was waiting on an answer to the original question so hadn't pressed it).
 
I was totally unfamiliar with the 5670 and using them with adapters in place of 6922's. What convinced me to try them was TD's sale advert and this article they linked to extolling their virtues as a 6922 replacement:

 
As @TriodeLuvr pointed out the differences between a 5670 and 6922 are significant, and it is only their mu (gain) number that is close. In some circuits a 5670 may substitute satisfactorily for a 6922, but that depends on the circuit and will never be optimal. In fine amplifiers like your Sonic Frontiers I would be surprised if they will function correctly. There may be enthusiast who tried the substitute with acceptable results, in specific equipment, but I have not heard of such. TubeDepot’s promotion of their 5670W plus adapter as “good-sounding 6922/E88CC” may work in certain instances but the statement can not be generally true. So, my advice would be to return the lot for a refund in unused condition.

The question of why some 5670Ws test good (when tested as 5670W on your B&K 747 ) and others not indicates that TubeDepot did not test the tubes before selling them to you, and shipped some that may be gassy or may have another defect.
 
Last edited:
I'll just add that the plate curves are also very different, as would be expected with such different Gm.

6922: 150V anode / -2V grid = 30mA
5670: 150V anode / -2V grid = 8mA

Absolutely not a valid substitute. My suggestion is to send everything back without further delay.

Jack
 
oof, yeah thats nowhere even close to the same.

I really wonder where some of these tube sub suggestions come from. I know there are cases of tubes that are electrically the same with a different pinout but sometimes you see stuff like this which isn't even close.
 
I appreciate the help and info. I reached out again to TubeDepot and they are processing a full refund with a prepaid shipping label. They said that they have sold a lot of these with the adapter as 6922 replacements and they have not heard of any problems (other than the occasional bad tube).

They offered to replace whatever I needed but, based on the advice above and the weird test results, I am doing the refund. One factor in my decision was the strength of the tubes as 6922 substitutes. My tester has a meter scale of 0 to 120 with Bad being 0 to 52. I'm not a tube expert, but if they all test 15-30 in the low Bad range as 6922's already, wouldn't that be similar to using a bunch of worn out actual 6922's that also test bad? And then what happens as these tubes wear over time and their values droop even lower? Would it be like using a tube that tests at 5 to 10 in the Bad range? Would I be cranking my stereo to 100 just to get moderate volume levels? There may be a tube expert out there that has more info, but I haven't seen enough to date to be comfortable that they will work well even with good replacements. (Based on your input above, the info suggests the opposite conclusion.)

The final factor was the uncertainty on how they might affect my preamp and amps. I don't know how the 5670's differences would affect other components in the circuit. Would the differences be harder on nearby passive components like caps and resistors? I don't have a clue but your answers above made me worry about that. The amps and preamp have worked flawlessly for 23 years and I don't want to take a chance that it will cause strain on something else. I need a new repair project like a hole in my head and these Sonic Frontier units are not trivial beasts.

Thanks again for the help and insight. Also, kudos to TubeDepot for processing the refund. Michelle in particular was wonderful to work with.
 
Last edited:
they probably do function in some circuits, but functioning and functioning correctly aren't the same thing. Its the same line of thinking that says the 12A*7 family is all the same except for the gain. They definitely are not, though they share a pinout and in many audio circuits at least you can swap them and have it pass signal.

it probably won't let the smoke out or anything fatal but I wouldn't expect the tubes to bias properly, so they may be operating in a wierd spot in the curve that has them making extra distortion or something

looking at their curves though, they do appear to be pretty linear and might make a nice tube to build a preamp around.
 
gagdet73,

There is a company that is making audio products designed around the 5670. The company is iFi. One product is the Pro iDSD Signature: "a standalone unit, a DAC/preamp, a network streamer and a headphone amplifier."

iFi also maintains that the 5670 is a valid substitute for a 6922, stating on their website: "Q. Specifications aren’t the same for 5670 and 6922 vacuum tubes? A. This is true, however, real tubes also show a significant variance from notional parameters, the match between 6DJ8/6922/ECC88 and 5670/396A/2C51 is close enough that they work even in quite twitchy fixed bias circuits. The only parameter that is of very minor concern is the 5670’s 15% higher heater current consumption, (350mA vs 300mA). Actually, various datasheets for the 6DJ8/6922/ECC88 group of valves (including Sino/Sov copies) show heater current ratings of up to 365mA maximum. The 5670 is tightly specified for heater current (more so than the 6DJ8/6922/ECC88 group). Any amplifier/pre-amplifier/DAC using the 6DJ8/6922/ECC88 group tubes should provide enough heater current for the stated maximum and will work fine with the 5670."

TD had pointed to an iFi article about the tube and that's what lead me to buy them.
 
This about the same as replacing 6DJ8 / 6922 with Russian 6N1P - no adapter needed, but half the transconductance. Not a real substitute, but will work acceptably in some circuits.
 
Did you try to test them with 5670 settings on your tester? They are probably good tubes, just not a sub for 6922.
 
This about the same as replacing 6DJ8 / 6922 with Russian 6N1P - no adapter needed, but half the transconductance. Not a real substitute, but will work acceptably in some circuits.

So question for you, what effect will "half the transconductance" have in its performance in an amp or preamp? Is ti going to manifest like a "weaker testing tube-? For instance, they all tested "Bad" as 6922's with the adapter (e.g. 30 on my B&K747). Does that correlate with the lower transconductance? Would that also correlate to performing like a worn out 6922 that also tests at 30 on my tester?
 
Did you try to test them with 5670 settings on your tester? They are probably good tubes, just not a sub for 6922.
Yes, I did. My first call to TD was because they tested bad as 6922's with the adapter and trying to get info to explain that. While I was waiting for them to get back with an answer, I tested them last Sunday as 5670s. I was getting odd results. The "Pop and Drop" testing behavior I described in my first post happened with more than half of the tubes. Tested as 5670's. Some tested normally but I got the weird results with a lot of them. I did a lot of retesting, too. I even did some control samples with known good and weak 6922s and 12AX7s. My tester performed normally with the control samples of my own tubes.

(A handful of the 5670s tested bad in one section or another. Three others failed both shorts and grid emissions tests and I stopped there. One was obviously cracked and had gone white. My comments about testing were on the remainder of the tubes.)
 
"...real tubes also show a significant variance from notional parameters,"
That's a poor excuse for lying about these being an across-the-board substitute for the 6922. The first place I ever saw the 5670 was a Chinese something or other on eBay. After I looked at the curves, it was clear the only reason they used this type was its abundance and extremely low cost. That was well before the Russian invasion of Ukraine, and there were many readily available (although more costly) tube types better suited to their use.
TD had pointed to an iFi article about the tube and that's what lead me to buy them.
I did read that after you posted the link. Pure marketing, absolute garbage, simply designed to sell tubes bought cheap at a profit. Could be wrong, but I think TD is the same seller that used to claim the 6P3S-E is equivalent to the 6L6GC. The 6P3S-E is an excellent tube, but not the equal of the GC.
I really wonder where some of these tube sub suggestions come from.
From sellers with no ethics.

Jack
 
The B&K 747 measures transconductance (at a fixed operating point, not under data sheet conditions), so I'd expect 5670s to measure lower than 6922s. Doesn't mean it will work like a weak 6922, since transconductance is only one parameter of many.

Tube Depot also sells 6SN7 to 6DJ8 adapters - those tubes are really different, but might work acceptably in some circuits (and not well in others...).

I did buy a pair of 6GH8 to 7199 adapters from them - in the one case I tried, the operating voltages were identical with either tube, so I'm satisfied with that swap.
 
This about the same as replacing 6DJ8 / 6922 with Russian 6N1P - no adapter needed, but half the transconductance. Not a real substitute, but will work acceptably in some circuits.

Yeah, you can usually readily use the 6n23P- it's dose to identical to the 6922, except for higher power handling and very slightly higher heater current- but the 6n1P just isn't going to bias up correctly in some circuits made for 6922/6DJ8.

Regards,
Gordon.
 
Back
Top Bottom