Grounding?

[Blair]

Thanks Ty,

I found those threads last night and replaced the 4.7K resistors with 2K resistors. It was helpful. but I will put 1Ks in today. I bought this quad as a perfect matched quad and they have less than 100 hours on them.

I may have been biasing them wrong all along though, so maybe you could clarify.

1) is the bias voltage of .7v more important than the -58v being dead on?

2)Do you typically balance the two until they have the same bias voltage with the 500K pot and then use the 10K to get them to the .7v level regardless of the voltage present at points G & F?

Thanks,

Blair

[Blair]

OK,

I swapped the tubes out like you suggested and adjusted the 500k pot to where they were balanced. Then adjusted the 10K pot until they were at .65v. One tube still is very slightly glowing on the plate, but it may not be obvious when the amp is turned over and light is present. I backed it off to .65v after noticing this. It seemed to help a little. These are those same SED tubes you referenced in the links this morning. Should I back them off to .60v?

All the voltages at each channel's points G anf F are within 3-4v of each other, but closer to -65-70v than -58. Is this OK?

Thanks again,

Blair

[Ty_Bower]

1) is the bias voltage of .7v more important than the -58v being dead on?

Absolutely. The 700 mV drop across the 10 ohm resistor gives a measurement of the idle current through the tube. I=V/R, so 700 millivolts divided by 10 ohms equals 70 mA idle current.

The -58v bias voltage is just the means to set the idle current where you want it. If the idle current is too high, then you need to apply a greater negative bias voltage. The electrons are cooked off the cathode by the hot filament. They want to fly through the vacuum to the plate because they are attracted to the large positive voltage there. If left to their own, you'd have an enormous current flowing through the tube (at least until it melted). Putting a negative voltage on the grid "scares" the electrons back towards the cathode and doesn't let so many of them get to the plate.

2)Do you typically balance the two until they have the same bias voltage with the 500K pot and then use the 10K to get them to the .7v level regardless of the voltage present at points G & F?

When starting with a brand new set of tubes, I'll adjust the 500k multi-turn trimpot on the Poseidon board until it is right in the middle. This assumes the KT88 are reasonably matched. Then I turn the 10K pot fully counterclockwise, which on a Mark III would set the maximum negative bias voltage. At this point you could power it up with the rectifier pulled out and verify with a meter that there was roughly -65 volts DC at pin 5 or 6 of the power tube sockets.

With the tubes and the rectifier in, power it up and monitor the voltage drop across the 10 ohm sense resistor. Fiddle the 500K trim pot until both sides show an equal voltage across their 10 ohm resistors (from the bias test point to ground). You can even stick one meter probe in each bias test point and measure the differential between them. You want the differential to be zero - the power tubes are balanced at this point.

After than, I'll slowly turn the main 10k bias pot clockwise and check each side to see that it goes to 70 mA (700 mV). It may be necessary to go back to the 500k trimpot again and re-balance them as you bring the idle current up.

Keep in mind the 70 mA specification is for KT88 with about 500 volts at the plate. 70 mA times 500 volts equals 35 watts, which ought to be within the dissipation limits for the KT88. If your plate voltage is substantially higher you will need to reduce the idle current to keep the dissipation within limits.

[Ty_Bower]

One tube still is very slightly glowing on the plate, but it may not be obvious when the amp is turned over and light is present. I backed it off to .65v after noticing this. It seemed to help a little. These are those same SED tubes you referenced in the links this morning. Should I back them off to .60v?

All the voltages at each channel's points G anf F are within 3-4v of each other, but closer to -65-70v than -58. Is this OK?

Again, the actual bias voltage required to achieve the desired idle current is less important than the idle current itself. If you need to run closer to -70v bias to keep the tubes idling properly, it tells me you either have relatively high plate voltage or relatively "hot" tubes. I wouldn't view either of those conditions as a problem provided you keep the overall dissipation of the tube below its limit.

If you can still see red glow, I'd try to lower the idle current further. I don't fully understand the chemistry of anodes, but I've been led to believe that if they are at the point where they are hot enough to emit visible light (i.e., glow red) then they are boiling out dissolved gasses in the metal. If left uncorrected for long periods of time this will erode the getter and shorten the life of the tube.

[Blair]

Thanks for the explanation Ty,

I backed the tubes all down to .60v which killed any redness on the plates. Again, the voltages are a little different on points G and F, but since the idle current is OK, all should be well if I am reading correctly. I also think I made a mistake by making the test points in series with G & F instead of putting them where the 10 ohm resistors are so I can actually see idle current. Either way, all seems to be good now and I just need to try and work on the slight hum issue. The red glowing tubes was more of a concern than the hum last night.

Blair

[Blair]

Now that I have the red plate resolved thanks to Ty and Tom, I would like to address my original post. Tom, I unhooked the input at the boards and shorted them, but it did nothing to the hum. Could someone explain the way they grounded their amps?

Here are a few differences in my amp vs. the mono poseidon version.

I do not have a CT tranny. It is a Toroid using two FWBs. I read all the different grounding articles typically referenced here as well as a few saved emails, but this does not follow that protocol. Here is how my amp is grounded.

The top plate is directly grounded from the IEC. The Driver boards are fed a ground through the RCA grounds, then branch to the PS via the point C on the boards. One thing I thought of that may cause an issue is that bother boards are fed from two different RCAs, and I should possibly run a jumper from one board to the other to try to eliminate a "loop" of sorts,but everything else jumpers to the PS ground thogh there is not much else. The bias supply and the driver PS caps.

Any ideas?

Thanks,

Blair

[patrickn]

I used the following site to learn about grounding and used it as a reference for grounding my amps:
http://www.aikenamps.com/StarGround.html
There's tons of great info on this site even though much of it pertains to guitar amps.

My Posiedon uses something like a bus ground because of the SDS board.
The items connected to the SDS board ground in order are:
1. The HV center tap
2. The bias supply and filament center tap
3. The KT88 cathode resistors
4. The posiedon board, negative feedback, and input grounds.

With the exception of the bias and filament circuits, everything is ordered from most to least amount of current.

You can see a picture here:
http://diytube.com/phpBB2/album_pic.php?pic_id=57

Hopefully this helps.
-Patrick

[Blair]

I used the following site to learn about grounding and used it as a reference for grounding my amps:
http://www.aikenamps.com/StarGround.html
There's tons of great info on this site even though much of it pertains to guitar amps.

My Posiedon uses something like a bus ground because of the SDS board.
The items connected to the SDS board ground in order are:
1. The HV center tap
2. The bias supply and filament center tap
3. The KT88 cathode resistors
4. The posiedon board, negative feedback, and input grounds.

With the exception of the bias and filament circuits, everything is ordered from most to least amount of current.

You can see a picture here:
http://diytube.com/phpBB2/album_pic.php?pic_id=57

Hopefully this helps.
-Patrick

I have the negative side of my FWB grounded, my Bias is grounded to the same point/bus, and my cathode resistors are grounded to the bus as well. There is no 6.3v CT or HV taps on this tranny, though I could make one for the 6.3v filaments. My input grounds are grounded through the boards, and the boards run straight to the bus.

When you say you grounded your negative feedback, do you mean you ran a wire to the ground bus from point I on your boards?

Thanks for any clarification.

Blair

[patrickn]

No, the negative feedback is grounded by a wire from the negative output terminal to ground. Point I needs to go to the 16ohm output terminal.

[Blair]

Thanks,

That is the way I have it wired. I'll try other means.

Blair

[TomMcNally]

Blair -

You might want to experiment with a center-tap on the 6.3 v
winding - either real or artificial using a pair of 100 resistors
from the filaments to ground.

[Ty_Bower]

The top plate is directly grounded from the IEC.

While I would never recommend you lift the green safety ground wire from the power cord, what happens if you lift the safety ground?

[Blair]

Blair -

You might want to experiment with a center-tap on the 6.3 v
winding - either real or artificial using a pair of 100 resistors
from the filaments to ground.

Do you ever have the same avatar Yellow_Light_Colorz_PDT_08

I'll give that a try. I need a day or two to let my back heal from flipping this thing over. It is a bear! I would assume 1-2W resistors will work?

Blair

[Blair]

The top plate is directly grounded from the IEC.

While I would never recommend you lift the green safety ground wire from the power cord, what happens if you lift the safety ground?

I'll try that, but like I told Tom, I just flipped this thing over for the last time today. It is killing me.

If I remember though, I did this and nothing happened. I'll try it again though.

Thanks,

Blair

[Ty_Bower]

Do you ever have the same avatar?

No, Tom builds way too many amps. Yellow_Light_Colorz_PDT_16