ST-120 Rectifier Puzzler…

jdm wrote:...I still am not sure why the new GZ34 tubes worked for 2 years and now only the NOS works... Is the problem current back flow or total current demand?

The problem usually can be traced to high AC line voltage.

At the minimum you should have a AC line monitor where you can check the line voltage before you turn on your tube equipment.

I use a Furman Elite-15 line conditioner, it holds the line current at 118 volts. What monitor do you recommend to monitor the Furman?

Wow any cheaper DIY alternatives to Furman Elite-15 line conditioner?

It won't regulate power, but a Kill-A-Watt meter is your friend when it comes to keeping an eye on your power situation.



Just plug it in between the amp and wall and let it go for a while. It'll give you averages over extended periods, as well as spikes and dips. It'll also let you monitor power draw, which is real handy. Neet little tool.

$22 at Amazon ...

Those thermistors are funny beasts as I tried them at work in an inrush situation with mixed results.  Comparing the various CL types as their max. current capacity goes up the cold resistance is much lower.  The CL90 cold resistance would be ideal at 120 ohms but its maximum current rating is 2A.  Other CL types have way less cold resistance so wouldn't be as effective at preventing inrush.  Plus with a TDR board the amp heaters alone pull about 20-25% the CL90 rated operational current before the relay kicks in to supply B+ voltage.  If the thermistor is installed on the AC input the CL90's resistance is 7.8 ohms at 25% power.  So I am not sure it would have that much effect in this circuit position when the TDR kicks in.  The AC line thermistor would be pre-warmed when needed most in this case.  

Also you want to operate a thermistor at 75 - 100% rated current as they have least resistance when fully warmed up.  If the CL90s are installed on the rectifier HV AC inputs the B+ would pull way less than the CL90 rated current.   When operating thermistors in the 0 - 25% rated current range the resistance varies wildly on slight current demand changes.  This varying thermistor resistance would modulate the B+ voltage with the amp's current demand.

And thermistors are not very effective on short term power failures since they are already warmed up when power is returned.  Anyway, that is my experience with these critters.  I never could find one that satisfied all my requirements.   

Thermistor Datasheet
Edit - correct some typos...

Agreed - there is no one stop solution.

That said, I feel a lot better going with -

- a thermister (and I still think the CL90 is our best bet due to the warmup characteristics even with it's lower amp rating
- an inline GFCI to prevent the hot starts you mentioned on a power blip
- the rectifier diode mod
- a properly tuned amp

Worst case scenario, a variac run inline to drop the power if it's high average. I did find that it can take quite a while for those to stabilize, so best bet would be to leave it on constantly. And worth repeating, if it's just a couple points high, the thermister once again never goes fully open, and will drop the voltage some, so you may just get away with that if it's on the low side of high.

Still on the fence about the bottle vs SS rectifier thing, as there's the issue of what they sound like. To each there own there, and if a bottle sounds better to you personally, well ... that's the most important part of the equation.

I still don't have a clue why the OP's having the problem he does, unless pure dumb luck pointed him to a bad batch of tubes. Even then, first step is to use something like the Kill-A-Watt to see exactly what the incoming power looks like. Maybe an upgrade in the neighborhood like I recently had when they switched out from 4k to 14k lines ...

Maybe out amps should be slightly redesigned ;
- at power on the rectifier is disconnected except for a resistor that allows slow charge of caps, slow enough so charging time is app. 10s
-after charge up delay the resistor is bypassed .
-any power outage will restart the powerup sequence, thus protecting from power blinks.

The implementation could be relay+resistor+control or it could be part of a regulated PSU whatever
is convenient. Note however that regulating B+ also means that bias must be regulated.
Whatever method used a tube rectifier would have a much easier life.

How is the delay relay that tubes4hifi designed ? Does it the above ?

I believe it does ... a logic controlled soft start and all.

I also understand you can adjust the soft start timing with a simple resistor swap.

Pretty neat really - I avoided one because of the transistor inna tube omygawd factor, but come to find out, the new VTA board already has a couple anyway. Not to mention I just added a 24v power supply board to drive the annular light on my power switch. So now it's ... what's a couple more transistors?  

PeterH has a pretty good idea.  Resistance has to be added somewhere to control inrush current.  But when the caps are sufficiently charged I would want the resistance to disappear otherwise the power supply would be sugglish in response to current demands.

So I see two proposals.. Add resistance to bypass each leg of the HV on the original TDR relay circuit.  The resistance would be the current limiter as the rectifier warms and starts slowly charging the caps.  Then when the relay board times out the resistors are each shorted and the HV from the transformer is connected directly to the rectifier (or diodes) as it is now.  The original circuit features remain intack.  The diode mod would be a must in this circuit for safe protection.

Or if you say no B+ until the heaters have warmed add a 2nd relay board that times out after the original or 1st relay design.  The 2nd board has B+ connected and a resistor also bypasses the relay to slowly charge the caps when the 1st or HV relay kicks in.  After the caps charged the 2nd relay times out and shorts the current limit resistor.  And again the original power supply circuit remains whole.   Perhaps the 2nd relay board can piggyback on top of the original.

Just some ideas.  Thanks...

BTW, no transistors on the VTA driver boards, but their is an SS (OHG!!) CCS, which is used purely as a constant current source and is not in the signal path.

I think you guys are getting warm.... If I remove the relay board that would eliminate the black box that I do not have the tools to check. Is it worth the time to remove the relay board to see if the New GZ34's operate as before. Do we no the failure rate of the parts on that board?

I am currently running a 1950's RCA 5U4GB and it sounds as good as my 58' Mullard GZ34. (Thanks Bob)

Something changed recently and the amp rejects the new GZ's at normal line voltage. I measured the line volts with a good multi meter at the same time I was testing the GZ's. The line volts were stable until Randy Travis was admitted to the Hospital down the street. That could be a factor!

tubes4hifi wrote:BTW, no transistors on the VTA driver boards, but their is an SS (OHG!!) CCS, which is used purely as a constant current source and is not in the signal path.

Technically, the CCS used is a JFET? Which would make it at least kissing cousins with the classic transistor, AND/NOR/OR maybe more appropriately the bastard child of a CCD ...

I will withhold the neener neener bit until clarity is achieved.  

The line volts were stable until Randy Travis was admitted to the Hospital down the street. That could be a factor!

There you have it!  Case solved, move on to the next issue...  Poor old Randy though is in bad shape, wish him well!

Adding a constant current source to the B+ sounds like a great idea but no simple one-chip solutions exist that operate at 500V DC (that I know).  My issue with the thermistors was solved using a hot-swap controller chip that employs a current sense resistor to manage the current limit threshold.  The chip then drives a large MosFET accordingly to slowly bring up voltage.  Unfortunately, hot-swap controllers are only good to about +48V.  

If you want to operate without the TDR delay (as a test) just pull the capacitor on pin 6 of the TDR timer chip.  This cap is part of the RC time constant circuit that controls the amount of the delay.  You would then have essentially no time delay and the relay will engage immediately at power on.   This seems easier if you have thoughts of rewiring around it.   I'll leave it to experts whether this is a good idea or not.

The rectifier problem is related to the optional transistor relay board. I bi-passed the red transformer wires direct to the rectifier and my St-120 came back to life. I assume the relay or other part(s) on the board failed. I think at this point taking out the relay board is best for my St-120. Has anyone had this problem with the optional relay?

The diode mod looks like a must, is this mod part of the St-120 kit now?

Thanks Bob, you have been a great help anytime I have had a question or concern. As you know, I only know enough about tube electronics to get myself in trouble. I do know that my ST-120 sounds as good as any I have listen to at reasonable volume levels.

I will send you the relay board for your review. I may just go without at this point. I will add the diode mod when I pull the board.

Thanks for you help.