KT88 and grid leak/bias feed resistors

[Ed_Chambley]

Kt88's usually have 100k bias resistors for a pair.  I have seen 150k and 220k as well.  I am wondering what is the problem with using 220k since my current build uses a LARMAR Master which is just over 225k.

Other than being rough on the tubes, what other problems could this cause?  Would it be best to place a resistor across 1 and 3 of each the dual pot to get it closer to 100K or will the resistor cause havoc with the bias being connected to lug 3 on both pots with the 2m2 resistors on lug 2 and 3 of each pot?

I have the amp running and the bias trim has a great sweep, but I wanted to check to see.  I am not concerned with the bias running away and red plating as my startup and testing of the amp is stable.  Cutting to 100K loses a significant amount of gain which I would like to keep, but I don't want to wrap it up if this is a problem waiting.  If necessary, I can get a dual 100k pot, but I have seen 6550 and Kt88 running for years with the 220k.

[Willabe]

Hi Ed,

Glad your new build is working out so well!

I'll take a swing at your question about the grid bleed R's.

If I'm understanding you correctly about the 2 100K R's, they are not bias R's. They would be there if the PA stage was K bias too.

If you look in any tube manual you will see that for any power tube, there is a max rating for grid resistance to ground. This includes the total resistance to ground, ie, grid stoppers, which is noted. It will be spec'd for both K bias and grid bias, although a different value.

IIRC, the reason is, if the power tube is gassy the limit of a lower R to ground provides a low resistance (and low Z?) path to ground to keep the bias stable. Because if the tube is gassy it can mess with the bias.

Ken Fisher wrote that he didn't trust new production tubes on certain amps for this reason, so he recommended lowering the schemo value on classic amps, like Marshalls with EL 34's, from 220k to 100K.

With that being said, IIRC, PRR, HBP and others here (and else where) have pointed out that many production amps have cheated/raised the tube manuals ratings with no problems through the years/decades.


                 
                   Brad             

[Ed_Chambley]

That is correct, not the 82K/100K plate resistors, but the 220k grid that connect to the PI coupling caps.  The ones you remove to install the master.  I have heard them called different things, but grid bleeder is fine.

I am going to run it a little longer and check a few more tubes.  It seems fine, but it is a common thing to do when changing a Marshall from EL34 to 6550.

Are tubes the only potential problem? 

[jjasilli]

It depends on the circuit designed around the tubes, which affects output power & THD.  Typical circuits, one with 220K the other with 100K resistors, are shown in this tube chart:  http://www.mif.pg.gda.pl/homepages/frank/sheets/084/k/KT88_GEC.pdf

For guitar amps, THD is less of a factor or may even be desired.

+1 to Willabe that for every tube type there is a max value of G1's resistance to ground.

[Ritchie200]

Hey Ed,

How hot are you running these?  Are they KT88 or KT120?  Ritchie's amp tech bumped the grid up from 220 to 10watt 1K on his Majors.  But then he had them biased for 240+ watts.  Much more stable.

We need pics of this beast!

Jim

Wait....Are you talking about the screens?

[tubeswell]

The grid is in relatively close proximity to the (heated) cathode, therefore the grid has a tendency to warm up a bit itself, and can emit a few electrons. This would otherwise cause the grid to become charged, were it not for the grid leak, which allows these electrons to be replenished from the ground return, thus keeping the grid at the desired grid-to-cathode voltage. More powerful tubes have a higher propensity for the grid to warm up, so you want a smaller grid leak resistance (all other things being equal) to ensure the grid remains at its desired potential.

The amount of grid leak resistance is a bit of a balancing act, because the lower the resistance, the higher the AC load on the preceding (driving) stage. With this in mind, KT88 amps can benefit from a CF buffer stage between the driver stage and the KT88 grid to counter the lowish grid leak resistance.

[Ed_Chambley]

Hey Ed,

How hot are you running these?  Are they KT88 or KT120?  Ritchie's amp tech bumped the grid up from 220 to 10watt 1K on his Majors.  But then he had them biased for 240+ watts.  Much more stable.

We need pics of this beast!

Jim

Wait....Are you talking about the screens?

I am using old Tung-sol 6550 for now.  I seem to be having filter problems.  I had a series set of caps blow and I replaced all 4 of them, 2 sets of series caps both 450v/100uf with 270k bleeders.  Brought the amp up to full voltages using a variac.  Ran it without any tubes for 30 hours.  Voltage rested at 565vDc on the plates tubes in and about the same on the screens.  Last night I installed the tubes and gave it a run around.  The amp sounds great, but I have a problem I cannot determine to source.  The last series cap in the power supply is getting hot and a resulting clicking sound.  Much hotter than the other 3 and i began to get noise, so I shut it down for the night.  Doesn't do it without tubes.

I will try again tonight, but I do not understand why I am having filter cap problems.  I mean if they are wired wrong the amp will not work and they will be destroyed as soon as they receive current.  I am trying, with no luck yet, to determine what is stressing the caps.  Since I do not know exactly how to convert the Route 66 to run Kt88 or Kt120, I am having to do research and testing and written revision to the schematic as I go.

What could cause caps to overheat?  I do have a dual cap can for the preamp which is 50uf/500v, so all positions are the same capacitance.  I am going to try tonight to reduce the preamp to a dual 32uf/500v and see if that changes anything, but I don't have high hopes since I have Marshalls built the same way.

Changes I have made so far: I have made a SS rectification.  From the Diode intersections I get a B+ of 603vDc.  This to a series cap set of 100uf/450vDc (50uf/900v series) to a choke (55 ohms, 1.5 henries, 10% tolerance, 2500 volt DC hipot tested, 200 ma DC) to another series 100uf/450v where the CT of the OT is connected.  A 75ohm dropping resistor connects to the next cap which gives me the proper voltage drop for the PI and also voltages are good on the EF86.

I had it as an inductor input and it worked well, but had a significant voltage drop.  May have to go back to it, not sure.

[sluckey]

A hot filter cap is a defective filter cap. Good filter caps do not get warm.

[Ed_Chambley]

A hot filter cap is a defective filter cap. Good filter caps do not get warm.

Yes, I plan to replace it, but I have had one of a set explode.  I replaced all with new F&T's and brought the voltage up over 30 hours incase they had been sitting for a while.  This is the second set I am having problems with.  Makes no sense to me.  Does it make sense that I could have 3 filter caps, 2 Illinois and one F&T go bad or be bad to begin with.  I am thinking something else is the problem.  Any suggestion as to what it could be?  I really do not want to give up on the idea of the high voltage as the amp sounds really really good.

What is crazy is I have built the exact amp and it is running at a Night Club almost every night for about 4 months.  Only difference is the OT is a Hammond 1650 I think.  Don't quote me on the number.  I made them at the same time and added the transformers later to the second.

I am attempting to draw a current schematic as an sch file, but the learning process is slow.  Also, I cannot seem to get the custom components to load on my windows 8 professional 64 bit laptop I have at home.  I will figure this out tho.

I can draw this SOB in about 30 minutes from scratch if I used Adobe Illustrator, but I am intent on learning Express.

[Willabe]

Also, I cannot seem to get the custom components to load on my windows 8 professional 64 bit laptop I have at home.  

Ed any sch schemo posted here can be opened and taken apart for custom parts.


              Brad    

[Ed_Chambley]

Also, I cannot seem to get the custom components to load on my windows 8 professional 64 bit laptop I have at home.  

Ed any sch schemo posted here can be opened and taken apart for custom parts.


              Brad    

I just figured out how to make custom components.  Group, and from the components choose make custom components, then name it.  I now know what is the problem with my new Windows 8 machine.  Custom components need to be placed into the Express component library in the user document folder and not the program folder.  This is different than any program I have used.  Usually if you place components in the program folder, they would be available to all users.  The file type is different in the program as opposed to to the library in the user document folder.

Now I need to figure out why filter caps are failing.

[jjasilli]

Something is causing your filter caps to fail.  It could be an issue with one of the things in the list below.  But maybe this problem is that inrush current & voltage surge is damaging the caps at turn-ON.  Maybe use a varistor or more complex soft start-up circuit.  Or stack the caps in filter stage 1 in series (or 4 caps in series-parallel) for a higher voltage rating.

From the www.geofex.com tube amp debug page on the topic:  Smoke or burning smell (which I think is close enough to your issue):
Failing/shorted output tube - this can overheat the output transformer and/or power tranformer. More rarely, it can also overheat the choke, but usually the transformers go first.
Improperly biased output tube.
Failing bias supply on fixed bias amplifiers
Failing cathode resistor bypass capacitor on cathode biased amps.
Failing/shorted rectifier tube (or solid state diodes - they do fail, if rarely) can overload the power transformer as well as killing the power filter capacitors by letting AC through. A failing filter cap or shorted output tube can pull so much current that it overloads the rectifier tube, too.
Failing power filter capacitor. These can sometimes get hot enough to literally explode or burn, as well as just quietly overloading the power supply and popping rectifiers and power transformers.
Failing power transformer.
Failing power filter capacitor
Choke with a "soft" short between winding and core
Failing output transformer

[Willabe]

I just figured out how to make custom components. 

 

[PRR]

> component library in the user document folder and not the program folder

Win7 introduced new folder conventions. Ordinary users have very little privilege in Program Files (makes it harder for malware to be installed). Since older (<=XP) programs expect to be able to do this, there's some slight-of-hand into the User folder. Of course this trick doesn't work for everybody (or every program).

[sluckey]

After reading your new thread about this amp I strongly suspect you have the series filter caps wired wrong. There's a very simple test you can do to prove/disprove my suspicions...

Your B+ from the rectifier diodes is approx. 600v. So each cap in a series pair should see 300v. Use your voltmeter to check the voltage across each cap in the first series pair. (Meter probes connected DIRECTLY ACROSS THE CAP LEADS). Do this test for the second series pair also. The voltage across any of these four single caps should be about 300v. If you find approx. 600v across any of these four caps then you likely have a wiring error.

I know you followed Doug's instructions about wiring two series caps on an 8 lug terminal strip. But you may have misinterpreted those instructions. If not, I'll just shut up about this. Do look at the attached pic though. And maybe you could post some hi rez pics of your filter caps if this is not the problem.

[Ed_Chambley]

After reading your new thread about this amp I strongly suspect you have the series filter caps wired wrong. There's a very simple test you can do to prove/disprove my suspicions...

Your B+ from the rectifier diodes is approx. 600v. So each cap in a series pair should see 300v. Use your voltmeter to check the voltage across each cap in the first series pair. (Meter probes connected DIRECTLY ACROSS THE CAP LEADS). Do this test for the second series pair also. The voltage across any of these four single caps should be about 300v. If you find approx. 600v across any of these four caps then you likely have a wiring error.

I know you followed Doug's instructions about wiring two series caps on an 8 lug terminal strip. But you may have misinterpreted those instructions. If not, I'll just shut up about this. Do look at the attached pic though. And maybe you could post some hi rez pics of your filter caps if this is not the problem.

While I would say I am certainly not above making this mistake, in this case I did not.  I have removed everything in the power supply and plan to rework it.  I will place photos later.  The diodes have failed.  Not sure if that happened first or when the first cap went.  In either case, I am building a cap board and getting some 3 amp diodes.  I know the diodes can only take a max of reverse voltage of 50 or so and I am not sure if a cap failing could do this, but both diodes measure infinity in either direction.  I wired them backwards and they light up a light.

I cannot make your measurements right now, but will to insure the new board is wired correctly.  Thank you for the input and testing method.  I really did not like the tag strip method because of where the caps must be in the layout.  In rethinking it, it will be best if I raise the caps a little since they are directly over the heater wires.

In retrospect, I think my inexperience in using SS rectification bit me somehow.  I know I should have checked the diodes when the first cap went, but I did not.  I will get the amp running stable as I seem to get determined with things like this.  Since it is a simple build, I am using the time to learn some additional things like using Express.  I had no trouble at all with Visio, but it seems as most here use Express and I just want to be one of the gang.

BTW, I do not want you to shut up about anything.  I am very type A and prone to mistakes because of it.  You can suggest anything as many times as you want.  I appreciate what you have taught me.

[PRR]

> I know the diodes can only take a max of reverse voltage of 50 or so

You can buy 1,000V diodes. In DIY quantity they are same-price as 50V diodes. You should only be using 1,000V parts (even in 12V systems, just because 1000V is so cheap by the bag-full).

However. Two 300V caps suggest you expect near 600V. In several rectifier systems, the diode reverse voltage is *twice* the DC you get. So you should be using 1,200V diodes. And you should always round-up 20%-50% for utility power bobble and line surge.

Another tidbit. BIG tube amps can over-current the 1A diodes. If total cathode current approaches 300mA, the 1A part can be in trouble (due to the spike-wave way we make DC). Four 6L6 or a hard-worked 6550 pair can be in this zone. (I got there with one 6550 once but it was an unusual setup.)

> getting some 3 amp diodes

Good to have and IIRC they can be had in 1,600V ratings which covers a lot of sky.

[Ed_Chambley]

> I know the diodes can only take a max of reverse voltage of 50 or so

You can buy 1,000V diodes. In DIY quantity they are same-price as 50V diodes. You should only be using 1,000V parts (even in 12V systems, just because 1000V is so cheap by the bag-full).

However. Two 300V caps suggest you expect near 600V. In several rectifier systems, the diode reverse voltage is *twice* the DC you get. So you should be using 1,200V diodes. And you should always round-up 20%-50% for utility power bobble and line surge.

Another tidbit. BIG tube amps can over-current the 1A diodes. If total cathode current approaches 300mA, the 1A part can be in trouble (due to the spike-wave way we make DC). Four 6L6 or a hard-worked 6550 pair can be in this zone. (I got there with one 6550 once but it was an unusual setup.)

> getting some 3 amp diodes

Good to have and IIRC they can be had in 1,600V ratings which covers a lot of sky.

If in series,  the voltage is doubled but not the amperage.  Is this correct.  I plan to rebuild it with 4 (2 sets in series) 1000v/3A diodes.  I used 1N4007 initially.  I plan to use 1N5408's.  With 2 in series with each HT I should be 2000v/3A.  Correct? This should handle the voltage and current, right?

Forward voltage is when the anode is more positive the the cathode.  Reverse is opposite.  How much reverse voltage does it take to send a diode into breakdown?  Can the first filter cap shorting cause reverse voltage on a diode?

[sluckey]

How much reverse voltage does it take to send a diode into breakdown?

The 1000V rating on the 1N4007 or 1N5408 IS the peak inverse voltage PIV rating. That's a tech way of saying those diodes have a 1000V reverse voltage rating. So for those diodes you're safe with a reverse voltage up to 1000 volts.

Can the first filter cap shorting cause reverse voltage on a diode?

No. In the case of your typical power supply rectifier circuit the reverse voltage is caused by the AC voltage from the PT. The PT AC voltage is positive for a half cycle then it reverses and becomes negative for the other half cycle. During the positive half cycle the diode is forward biased and acts like a closed switch to pass the positive voltage on to charge the filter caps. During the negative half cycle the diode is reverse biased and acts like an open switch preventing the negative voltage from passing thru to the filter caps.

For example, say your PT puts out 400-0-400. That's the RMS value and we are concerned with the peak voltage when talking about diode PIV ratings. So, that same PT puts out 566 volts peak. During the positive half cycle that 566v pulse passes thru the diode to charge the filter cap and during the negative half cycle that 566v pulse is blocked from reaching the filter cap. After several AC cycles the filter cap will be totally charged to that peak 566 volts. This is a no load condition. The cap will just continue to hold that positive 566 volt charge since there is no load current to bleed it down.

Now you might think that since the PT AC voltage is 566-0-566 peak that using a single 1000v PIV diode would be sufficient. But NOT! True, there is a negative 566v pulse on the anode whenever the PT AC voltage swings negative, but there is also positive 566v dc on the cathode. That means there is actually 1132 volts across the diode. And it is reverse polarity. The diode will likely fail. So, you put two diodes in series to take the PIV rating up to 2000v for adequate protection.

Fender used 3 diodes back in the sixties, but that's because he did not have 1000V diodes. Silicon diodes were fairly new back then and affordable consumer diodes had a typical 400v PIV rating.

[Ed_Chambley]

Understood Sluckey.  Thanks.

Here is one for you I do not understand.  When I copied the Fargen Old 800, the original had 4 diodes, 1000v/3A.  2 series from the HT to a single B+ tap.  The only thing different is he had the HT wiring connected to the intersection of the 2 diode like he was splitting them.

Diode-Wire-Diode.  What is the purpose of this arrangement.  When I copied it I did not do this.  I connected them like a Fender.

[sluckey]

Here is one for you I do not understand.  When I copied the Fargen Old 800, the original had 4 diodes, 1000v/3A.  2 series from the HT to a single B+ tap.  The only thing different is he had the HT wiring connected to the intersection of the 2 diode like he was splitting them.

You mean like the attached pic? The black circuit was drawn from your description. I added the green circuit because I believe that's the rest of the story.

[Ed_Chambley]

Here is one for you I do not understand.  When I copied the Fargen Old 800, the original had 4 diodes, 1000v/3A.  2 series from the HT to a single B+ tap.  The only thing different is he had the HT wiring connected to the intersection of the 2 diode like he was splitting them.

You mean like the attached pic? The black circuit was drawn from your description. I added the green circuit because I believe that's the rest of the story.

I see.  Makes sense to me now.

[PRR]

> Can the first filter cap shorting cause reverse voltage on a diode?

Reduces the reverse-voltage stress.

But the forward-current stress becomes "infinite".

[Ed_Chambley]

After reading your new thread about this amp I strongly suspect you have the series filter caps wired wrong. There's a very simple test you can do to prove/disprove my suspicions...

Your B+ from the rectifier diodes is approx. 600v. So each cap in a series pair should see 300v. Use your voltmeter to check the voltage across each cap in the first series pair. (Meter probes connected DIRECTLY ACROSS THE CAP LEADS). Do this test for the second series pair also. The voltage across any of these four single caps should be about 300v. If you find approx. 600v across any of these four caps then you likely have a wiring error.

I know you followed Doug's instructions about wiring two series caps on an 8 lug terminal strip. But you may have misinterpreted those instructions. If not, I'll just shut up about this. Do look at the attached pic though. And maybe you could post some hi rez pics of your filter caps if this is not the problem.

Sluckey, as I bring it up with a variac, I am getting equal voltage across each cap.  It takes a few minutes to balance, but seems fine.  Currently 120 vdc across each cap.  We will see when I bring it up to 600 or so.