Considerations when replacing GZ34 with 1N4007's

[nevetslab]

First I'll ask if anyone thinks there would be a benefit to having a standby switch so you can at least get the heaters warmed up before hitting the plates with B+ ?
The biggest issue I see with SS rectifier is the price of 300B tubes. If the slow ramp-up of tube rectifier B+ is helping power tube life, you really need to emulate it.

Now, to further flog the mort cheval, and incur the wrath of a sun god worshipper , why would they use a different vendor socket for the rectifier and preamp tubes? They're all octos. That's just bizarre. So it's not this particular 5AR4, and either of the 6SL7's have the same issue in the recto socket? The pins in the socket are still locked in place and have not broken free and got pushed out of place? I wonder if someone had done something funny (mod) with this unit to use some other rectifier or something.

I know we're all hindered from not seeing the physical photos of the Rectifier Tube socket, as well as the two Preamp tube sockets for the 6SL7 tubes. I assume like the rest of us have, that all three tube sockets are the same. Or is that NOT the case? I've not encountered LONG-PIN rectifier tubes, as what we're being told appears to be what's needed in this case. From the photo in the manual, all looks normal, as far as the three octal tubes in the photo. We really can't see any detail to suggest anything odd. No special added base to replace bad pins on the rectifier tube pin base is being used.

I know this is going out on a limb, but, having had to re-solder octal tube pins periodically in hopes to restore the operational state of an output tube in a set of power tubes, many times with success, that lends credence to perhaps scavenging a set of pins from some bad octal tubes, or similar source, heating up the pins in the 5AR4, inserting a suitable contact lead, such as 20AWG buss wire, and trying to add additional pin length to what's on the tube. Obviously the full length of the second set of pins would be too long, but with careful slicing of pins with a Dremel fitted with carbide discs for slicing, one could cobble the extension of the pins on a good 5AR4 so the longer pins DO engage the socket terminals. Hard to really see if this would work, NOT seeing the physical tube socket as you have there. I too am bewildered with this problem-induced tube socket as has rendered this amp non-functional.

Granted, adding a Standby Switch on the back of the amp out of sight would be a solution to the warm-up time, but would work, if there's no other solution. Creating a slow-ramp circuit to use rectifier diodes would be a lot of work.

[glebert]

Is John L. Lumley alive? Can he be contacted and asked what the heck is up with this socket?

[Mick Bailey]

I've had sockets where I've fitted the tube and had to push the pins from behind and they've moved a fair bit to give full engagement. When new the pins would have been fully secured though. I noticed that the Chinese wafer-type of transmitting tube sockets can be borderline - either the ceramic is too thick (or maybe the surface glaze) or the individual pin sockets are positioned too far off the underside.

My own view is I'm reluctant to make a design change to an amp to get round a problem; the amp is tube rectified and this is not the issue, but the socket is. Converting the amp to SS rectification doesn't fix the problem - it sidesteps it through re-design and gives rise to other concerns. Just my own personal view.

[g1]

First I'll ask if anyone thinks there would be a benefit to having a standby switch so you can at least get the heaters warmed up before hitting the plates with B+ ?
The biggest issue I see with SS rectifier is the price of 300B tubes. If the slow ramp-up of tube rectifier B+ is helping power tube life, you really need to emulate it.

Now, to further flog the mort cheval, and incur the wrath of a sun god worshipper , why would they use a different vendor socket for the rectifier and preamp tubes? They're all octos. That's just bizarre. So it's not this particular 5AR4, and either of the 6SL7's have the same issue in the recto socket? The pins in the socket are still locked in place and have not broken free and got pushed out of place? I wonder if someone had done something funny (mod) with this unit to use some other rectifier or something.

[Helmholtz]

Now what really makes me wonder, is the manual's instructions about adjusting the hum balance.

Using a DC current meter at the speaker output - what??

[Helmholtz]

Regarding sag:

It's a class A amp. Class A means constant power consumption at all signal levels - so no dynamic sag. Sag isn't desirable for HiFi amps anyway as it reduces dynamics.

Dropping the extra voltage will produce some heat - but not more than the GZ34 did.

Diode switching noise will most probably be no problem because of the 2-step choke filtering.

Fast voltage ramp-up might show as a turn-on thump. Maybe not.

[Mr_bibbles]

Being a SET amp you don't need to worry too much about sag, but the over-voltage that a straight SS rectifier gives needs addressing. Measuring the current draw should enable you to calculate an appropriate resistor, but this will produce additional heat inside the enclosure. Another way is to use a MOSFET dropper, but again the excess voltage is dissipated as heat - though in most cases the chassis can be used as a heatsink. Another possibility is to use a Zener. The slow startup is more difficult to emulate because the warm-up time of a GZ34 is much longer than many thermistors. Some browsing of spec sheets needed. You may need to pick your diodes to prevent switching noise - maybe some ultra fast devices.

Thanks so much Mick this is exactly the kind of thing I was asking for! I'll read into all of these options. As for using a resistor, the enclosure has PLENTY of space and is hand wired PTP so dissipation shouldn't be an issue, I usually over-rate my droppers when repairing too.

Replacing the GZ34 with SS diodes should increase B+ by less than 20V. Might not be a problem at all.

If you want to drop the extra voltage, please post B+ and 300B cathode voltage for advice.

Thanks Heimholtz! That's pretty much what I was thinking, but I wanted to look at it from all angles since in case of potential collateral damage. I'll do a bunch of comparative tests from before and after if I do decide to go this route and if anything seems rotten I'll definitely check in.

All hail the sun god,

Mr_Bibbles

[Helmholtz]

Replacing the GZ34 with SS diodes should increase B+ by less than 20V. MIght not be a problem at all.

If you want to drop the extra voltage, please post B+ and 300B cathode voltage for advice.

[Mick Bailey]

Being a SET amp you don't need to worry too much about sag, but the over-voltage that a straight SS rectifier gives needs addressing. Measuring the current draw should enable you to calculate an appropriate resistor, but this will produce additional heat inside the enclosure. Another way is to use a MOSFET dropper, but again the excess voltage is dissipated as heat - though in most cases the chassis can be used as a heatsink. Another possibility is to use a Zener. The slow startup is more difficult to emulate because the warm-up time of a GZ34 is much longer than many thermistors. Some browsing of spec sheets needed. You may need to pick your diodes to prevent switching noise - maybe some ultra fast devices.

[Mr_bibbles]

Yes, yes I know it sounds like I'm crazy! I am not in the shop over the weekend and I will definitely get pictures when I can. It's always possible I have overlooked something, Especially when encountering something new so I will triple check everything.

The customer found this second hand and it has never once worked...

So lets say for now that using a tube in this socket is a heathen activity! A paganistic abomination that must be avoided at the cost of angering the sun god!

THEORETICALLY, if our ONLY option was to install a SS rectifier in the amp (which, again, I have never done and will probably never do again if possible)... what would we be watching out for? For the purposes of our own edification only, of course!...

*looks nervously at sun god statue*

[Enzo]

I can't help but fear we are missing part of the puzzle, since it seems so unlikely anyone would design an amp physically so that only a very specific specialized part was required

Photos of the offending socket would be SOOO helpful. Don't you have a friend, wife, neighbor, child, girlfriend, golfing buddy or SOMEONE who could snap a couple photos?

And just to be clear, your socket...if we could take JUST th actual socket part of the socket assembly out in our hand, and the tube shoved all the way to the hilt, the pins STILL wouldn't reach?

[Mr_bibbles]

Thanks for the reply, Mick. I would also wire in a different socket normally, but it's not at all like a normal octal you'd think of like I was saying above, and replacing the socket would result in rewiring half of the PSU so this seems like a better solution... I've never ever seen a socket like this before, and It's a matter of PIN DEPTH not tension. I don't know how to explain without photos, but so far pretty much everyone except Nevetslab is more interested in the goofy socket than my question! So let me just say...

I know this is an abnormal approach to a poorly fitting tube, and have considered the normal remedies like tensioning and socket replacement.

...Seriously!

You just have to take my word on it. I'll try and post photos afterward.

Any advice on the real nitty gritty of solid state rectification's global effects? That's really what I was hoping to learn, especially any indirect consequences. Thanks!

[Mick Bailey]

I would wire in a different socket. Did you take any measurements and compare the socket fitted to (say) a Belton?

[Mr_bibbles]

G1, it's definitely the 5AR4, and this socket is the only offender on the HiFi.

Pixel, that's a pretty great idea! Unfortunately since it's a HiFi all of the tubes are on display on a very nice finished chassis box... it might be a little unsightly, but I'll look around and see if I can find something that could work.

Enzo, the tube base fits flush with the socket base and not the chassis itself; it mates with the socket perfectly. THEN, while it's kissing the plastic, you can see the pins travel underneath and just barely even reach the actual socket pins. It's astounding.


Like you guys, my first thought was "There's no way this socket would work like this / I'm missing something / / etc etc " but after a thorough inspection (and a little laugh with the shop manager who took a peek) I'm considering the 1n4007 as the "least invasive" course of action. So let's say, for the sake of argument, the Space/Time Police have issued us a warning, and we are not allowed to install a new 5ar4 and are going to use 1n4007's to make a full wave rectifier. I'll try and get a picture going of the doom socket once my phone arrives, at least for the forum's entertainment.

SO

1. We know we'll have to test voltages to ensure no component is exposed to a voltage higher than it's rating

2. As per Nevetslab's post, the in rush "warm up" times for the B+ and other associated parts will be much faster. This might be totally fine since the rectifier is probably the most at risk for inrush damage, but a thermistor in series with the 5ar4 could help inrush times.

3. ????? Perhaps a dropping resistor to emulate the voltage drop over the 5ar4?


Perhaps I'm being overzealous here and a SS rectifier isn't going to be a big deal here, but I want to make sure first and foremost I couldn't damage this amp with some rookie mistake I'm missing.


PS Happy Fourth! I hope everyone is enjoying time with their family and friends today.

[Enzo]

Looking at the picture. Let me ask, is the socket merely mounted below the chassis surface, AND your tube won't fit through the hole in the chassis? And that is why it won't mate the socket?

If so, what are the odds of removing whatever spacers there are and mounting the socket flush with the chassis surface so the tube doesn't have to fit in the hole.