4xKT88 Design Help

[erock0138]

Hello, I'm in the process of planning to build a 4xkt88 bass amp. I've decided on using the hammond 278CX (400-0-400 @535mA). And hammond 120W UL output. It's Semi-based on the Sound City B150

My PT doesn't have a bias tap, so my question is about using a separate transformer for the bias supply. Does any one have a link or a calculation method on how to compute requirements for it? I've searched, but not had much luck.


Also, With my iron being only rated at 120W, Biasing at only 60% gives 109.2W output, has anyone had success with more output wattage with this transformer arrangement? Without stressing them of course.


Or, would another transformer set be better?

Thank you!

[pdf64]

There's no necessity for a bias winding / separate PT, eg http://bmamps.com/Schematics/marshal...__45w_1962.pdf
Though a low impedance supply would be better, considering the limit of 50k for grid leaks.
Bias supply is very low current; no benefit designing for more than a few mA.

How have you calculated the power output?

Try having a play with Interactive Valve Data Sheets

A HT of 565V seems way too high to keep plate dissipation reasonable.

[Leo_Gnardo]

I built an amp with this combination 1996. Worked very well but a couple of things to watch out for. First, there's not quite enough current to run 4 6550 filaments plus preamp tubes. After a while in service, I added a filament transformer so the PT wouldn't be stressed.

The OT required a special switch to allow 4, 8, and 16 ohm outputs. At first I did without the switch, not understanding what the proper wiring was. When I did understand, still did without the switch, just selected 4 and 8 ohm outputs because the owner of the amp had no 16 ohm cabs and probably wasn't going to use any such. Here's the birdseye lowdown on the OT secondaries: there are two 4 ohm windings. For 4 and 8, you wire the second 4 ohm parallel to the first 4 ohm. For 16, wire the second 4 ohm winding in series with the first 4 ohm. (In OT world, two 4 ohm windings in series yields 16 ohms.) See, what a headache. Recently one of our correspondents mentioned Mouser carries an alternate 120W Hammond OT with "E-Z wire" secondary, all 3 ordinary impedances without fancy switching or rewiring. I'm sure other suppliers carry or can order that OT for you.

For bias you can borrow a little current from one of the PT HV windings as Marshall often does, recommend 2W 150K resistor from one of the HV secondaries, you can half-wave rectify, filter and pass thru voltage divider for bias. And that 150K value isn't carved in stone either, you may need more or less depending how you design your bias voltage divider, one trimpot, one for each pair of tubes, or individual bias trim per tube.

Or, with a separate bias transformer, anything that will put you in the 80V range before the voltage divider/trimpot stage should work well.

I got nearly 150 watts at clip with a set of GE 6550A. After wearing them out over the course of years, the owner switched to KT88, still 120W plus, and bone shaking bottom end that just goes all the way down there, solid, of course with quality speakers. For those, good ol' Altec, JBL, EV would be good, and these days you can get a great bass tone from Eminence Kappa Pro for half the price of the old warhorses.

I'm sure others will have some good ideas, here they come... Good luck on your build and let us know how you're doing.

[g1]

Also, With my iron being only rated at 120W, Biasing at only 60% gives 109.2W output, has anyone had success with more output wattage with this transformer arrangement? Without stressing them of course.

The % bias is for idle conditions only and does not dictate the output power.

[erock0138]

Thank you for the replies, I did plan on using an external filament transformer for the heater supply, as I did notice the windings on the transfer however only rated for 6 A. I can use the back by using method, I have experienced it before, but didn't think about it for this build. 565 V does seem a little high to keep things within spec, so I will probably plan on using 500 V and about 455 for the screen grids. With a lower voltage needed, what would be a good way to soak up the rest from my supply? I'm going to be running for preamp tubes, 3 12ax7s and an ECC 81. Thank you for the link to the interactive load line plots are, I did not know that even existed. But is a very handy tool. I do hope to find the easy wire output transformer, is the other way describe definitely sounds like a pain.

[erock0138]

I have the eminence kappa 15" and I love it!! Great tone from it. Another thing I had noticed was the sound city had a seperate neg. feedback tap. How would it effect it to install feedback the traditional way? I haven't seen an output transformer like that before.

Thanks!!!

[Gaz]

Thank you for the replies, I did plan on using an external filament transformer for the heater supply, as I did notice the windings on the transfer however only rated for 6 A. I can use the back by using method, I have experienced it before, but didn't think about it for this build. 565 V does seem a little high to keep things within spec, so I will probably plan on using 500 V and about 455 for the screen grids. With a lower voltage needed, what would be a good way to soak up the rest from my supply? I'm going to be running for preamp tubes, 3 12ax7s and an ECC 81. Thank you for the link to the interactive load line plots are, I did not know that even existed. But is a very handy tool. I do hope to find the easy wire output transformer, is the other way describe definitely sounds like a pain.

No reasonable way to "soak up" ~65V in an amp like this. You really just need to get the right transformer. This one would be about 500V loaded, and no need for an extra filament supply.

ClassicTone 100W Project Style Universal Power Transformer

[Wombaticus]

No reasonable way to "soak up" ~65V in an amp like this. You really just need to get the right transformer. This one would be about 500V loaded, and no need for an extra filament supply.

A few thoughts here. First, the original poster indicated that they intended to use an ultralinear output transformer. As it happens, 560V on plate / 553V screens with 40% ultralinear taps is a datasheet operating point for a pair of fixed-biased KT-88s on the old Genelex datasheet. This will allegedly give 100W operating into a 4.5 kohm plate-to-plate load. So, for a quad of KT-88s operating into 2.25 kohm it should be pushing 200W. If that operating point turns out to be a little bit toasty for comfort or leads to undesirably short tube life, it would be relatively straightforward to use some 5W Zener diodes to drop the voltages a bit just to the screens only. The voltage on the plates isn't really that critical! If you wanted to drop 50V probably I would series a couple of 5W 25V devices. Each screen would be dropped separately. Note that using Zeners (or ultralinear mode for that matter) does not obviate the need for screen grid resistors, so you would still want at least 470R / 5W in series with the Zeners.

The other option for KT-88s of course is to use standard tetrode mode operation, but in that case the screens want to be only about half the plate voltage when the plate voltages are high. I recently completed an amp using a pair of KT-88s with plate at 610V and screens at 305V into a 5.5 kohm load. The power section is giving up about 80W clean, and 110W balls out. The amp is only a few months old, but so far the KT-88s show no sign of being unhappy at that operating point. Note that if you go for tetrode mode operation, it's quite convenient to use a Marshall-style rectifier into a totem-pole reservoir cap (it's that arrangement that looks like a FWB, but really isn't, with the center tap of the PT going to the junction of the caps), as you can then derive the screen voltage from the halfway point of the reservoir.

Either way, if you get into the business of paralleling KT-88s with high plate and screen voltages, it is probably highly advisable to include from the get-go a 10 ohm or so plate stop resistor in addition to the usual control grid and screen grid resistors, and it is likewise recommended to make the bias individually adjustable for the two halves of the push-pull. Oh, and if you're squeezing a lot of power out of KT-88s, pay attention to the maximum control grid resistance - don't abuse this parameter as much as tends to be the case in guitars amps. Right now, I'm myself getting pretty deeply into a project for a 6 x KT-88 bass amp -- the iron has been ordered, so no turning back now!

Cheers,

Paul Fawcett aka Wombat

[erock0138]

A few thoughts here. First, the original poster indicated that they intended to use an ultralinear output transformer. As it happens, 560V on plate / 553V screens with 40% ultralinear taps is a datasheet operating point for a pair of fixed-biased KT-88s on the old Genelex datasheet. This will allegedly give 100W operating into a 4.5 kohm plate-to-plate load. So, for a quad of KT-88s operating into 2.25 kohm it should be pushing 200W. If that operating point turns out to be a little bit toasty for comfort or leads to undesirably short tube life, it would be relatively straightforward to use some 5W Zener diodes to drop the voltages a bit just to the screens only. The voltage on the plates isn't really that critical! If you wanted to drop 50V probably I would series a couple of 5W 25V devices. Each screen would be dropped separately. Note that using Zeners (or ultralinear mode for that matter) does not obviate the need for screen grid resistors, so you would still want at least 470R / 5W in series with the Zeners.

Thank you! This was along the lines of what I was looking for. Your input is greatly appreciated. Do you know of any amps that implement the zeners on the screen grid supply? This is the route that I'm going to take.

Either way, if you get into the business of paralleling KT-88s with high plate and screen voltages, it is probably highly advisable to include from the get-go a 10 ohm or so plate stop resistor in addition to the usual control grid and screen grid resistors, and it is likewise recommended to make the bias individually adjustable for the two halves of the push-pull. Oh, and if you're squeezing a lot of power out of KT-88s, pay attention to the maximum control grid resistance - don't abuse this parameter as much as tends to be the case in guitars amps. Right now, I'm myself getting pretty deeply into a project for a 6 x KT-88 bass amp -- the iron has been ordered, so no turning back now!

Right on! That sounds awesome. Thank you again.

[Wombaticus]

Thank you! This was along the lines of what I was looking for. Your input is greatly appreciated. Do you know of any amps that implement the zeners on the screen grid supply? This is the route that I'm going to take.

You're not going to find this on a commercial amp, because -- as others suggested -- a commercial manufacturer would simply re-spec the transformer. We smaller fish don't have that luxury, so this trick can let you adjust a few tens of volts downwards on the screens. It's not really a suitable approach for more than that.

But let's clarify a couple of things... you said earlier that you were going for an ultralinear transformer, but in a later post I saw that you are proposing to have quite different plate and screen voltages. It sounds like you are not intending to run in ultralinear, but rather in tetrode mode (i.e. with separate screen and plate supplies)? This is a huge difference, so let's get this right. My comments before were predicated on the idea that you were actually running in ultralinear. If you are NOT going for ultralinear, I don't think the operating point you suggested in the later post (500V plate, 455 screens into a 1.9k anode-anode load) would be good at all... you would hit Vg0 waay below the knee, and this would be really tough on the screens. If my guess is right that your chosen PT would give you about 530V, then I think you would want something more in the range of 350V on the screens for that 1.9K load impedance... out of range for the Zener trick, and a bit of an inconvenience all the way round. It might be *slightly* more convenient if you found an OT around 2.25 kohms or so, then the rectifier arrangement I mentioned would naturally give you something in the 270V range which would be about right for screens at that load impedance.. But finding the right PT would then be a pain as well.

But..this is a bass amp, why not just go with ultralinear and be done?

[trobbins]

A little bit more complex is to use two rectified supply rails from the one PT. One supply is choke input filtered (ie. ~350VDC rail), and the other is the more well appreciated 550VDC from a capacitor input filter supply. The complexity is just a couple of extra diodes and a low current choke. The choke input supply is for low power preamps and screen, and so should end up being better filtered.

[erock0138]

But let's clarify a couple of things... you said earlier that you were going for an ultralinear transformer, but in a later post I saw that you are proposing to have quite different plate and screen voltages. It sounds like you are not intending to run in ultralinear, but rather in tetrode mode (i.e. with separate screen and plate supplies)? This is a huge difference, so let's get this right. My comments before were predicated on the idea that you were actually running in ultralinear. If you are NOT going for ultralinear, I don't think the operating point you suggested in the later post (500V plate, 455 screens into a 1.9k anode-anode load) would be good at all... you would hit Vg0 waay below the knee, and this would be really tough on the screens. If my guess is right that your chosen PT would give you about 530V, then I think you would want something more in the range of 350V on the screens for that 1.9K load impedance... out of range for the Zener trick, and a bit of an inconvenience all the way round. It might be *slightly* more convenient if you found an OT around 2.25 kohms or so, then the rectifier arrangement I mentioned would naturally give you something in the 270V range which would be about right for screens at that load impedance.. But finding the right PT would then be a pain as well.

The 500V Plate, and 455V screen voltages were from the tube data sheet that I had for Kt88's specifically for running in ultra linear operation, and they match the voltages of the amp that I am basing this off of. (Sound City Bass 150) But yes, I am planning on running ultra linear. Maybe I'm looking at this wrong? As you had mentioned before, the plate voltage I can keep high, just need to adjust for the screens, but with my output transformer being rated at 120W, I was wondering if I had to bias colder to stay within specs of the OT.


Thank you again!

But..this is a bass amp, why not just go with ultralinear and be done?

Yes! UL I believe is the way to go.

[pdf64]

The 500V Plate, and 455V screen voltages were from the tube data sheet that I had for Kt88's specifically for running in ultra linear operation, and they match the voltages of the amp that I am basing this off of. (Sound City Bass 150) But yes, I am planning on running ultra linear. Maybe I'm looking at this wrong? As you had mentioned before, the plate voltage I can keep high, just need to adjust for the screens...

With regular UL, the screen grids are supplied via the OT, ie closely similar VHT as the plates.
A separate screen supply is therefore not possible with such an OT.

Please provide links to the info sources you are referencing, ie KT88 data sheet, your B150 schmatic.
As the operating conditions you report don't seem to correlate with the available info, eg http://www.mif.pg.gda.pl/homepages/f...084/k/KT88.pdf http://soundcitysite.com/b150_schemo_right.jpg

[erock0138]

With regular UL, the screen grids are supplied via the OT, ie closely similar VHT as the plates.
A separate screen supply is therefore not possible with such an OT.

Please provide links to the info sources you are referencing, ie KT88 data sheet, your B150 schmatic.
As the operating conditions you report don't seem to correlate with the available info, eg http://www.mif.pg.gda.pl/homepages/f...084/k/KT88.pdf http://soundcitysite.com/b150_schemo_right.jpg

The links you have provided are the same I was referencing.
On the schematic, I know the writing is hard to identify, I interpereted it As the plates 495, screens 450.
Looking at the last load line graph also, I saw the 500v mark directly correlates with g2 voltage at 455. So that's where I was getting my basis off of, but I believe I see my error now.

[pdf64]

...On the schematic, I know the writing is hard to identify, I interpereted it As the plates 495, screens 450...

Sorry, just to confirm, can you see that there's nowhere between the OT CT and the tube screen grid that could have 'lost' 40V ?
Even 490V seems low; UL screen grid often idle at a higher voltage that their plate, especially beam tetrodes, which don't draw as much screen grid current as pentodes (I acknowledge that GEC describe them as beam pentodes).