zzz
500+ Head-Fier
- Joined
- Apr 3, 2002
- Posts
- 757
- Likes
- 66
My idea of spending this birthday was to mess around a certain HV amp. Turning 21 in Canada is not nearly as much of a deal as in the US, you know
.
It's been mostly a test-bed for Weebl. One channel only. The idea was to get some sound out of it and determine the temperature behaviour of the various part of the amp experimentally. Some do it with protoboards but that's for the weak.
Photo #1:
On the left one can see the vandalized KGSS (transformers disconnected and reordered around, +-15V power supply routed to the Weebl, oscillation-prevention caps cut out [to be used in Weebl]). Sitting on the box are +-400V/+580V power supply, filament transformer, input stage, and the new output stage. Dig the 3D (all over the place) component layout of the output stage!
Photo #2:
Focus on the tube/current source side of the output stage.
Transistors hanging in the air used to be mounted on the other side but that was only because their dissipation has been somewhat underestimated. The two without any heatsinks (pair of 2SA1967 supplying base current and setting offset for the anode current sources) could use a small stand-up or snap-on heatsink but do not seem to overheat without one either. Giant heatsinks for the anode current sources could safely be half this large as contrary to my expectations they only got mildly warm (~10W dissipation per each). 2SC3675 (pass transistor that creates -300V. ~6W dissipation) on two Aavid heatsinks joined together hanging nearby gets damn hot. But you see that large heatsink standing on the carpet? This transistor used to heat that one up just as well. Probably it just likes attention.
MOSFETs run warm but not hot. The ones driving the grid (2SJ79) are a little hotter than 2SK216s. They will get their small stand-ups eventually, but that'll be enough.
If crazy mounting (like this one) is used, the amp will oscillate in a very weird fashion and will refuse to listen to any adjustments. With Kevin's help, it has been mostly cured. Proper layout (due Christmas) should eliminate the remaining issues.
BTW, an easy way to see if any current is flowing through the tube is to look for the blue flame inside. Wicked stuff. Also, mono through Staxes (only 404 was used for testing) is not half-bad at all either.
It's been mostly a test-bed for Weebl. One channel only. The idea was to get some sound out of it and determine the temperature behaviour of the various part of the amp experimentally. Some do it with protoboards but that's for the weak.
Photo #1:
On the left one can see the vandalized KGSS (transformers disconnected and reordered around, +-15V power supply routed to the Weebl, oscillation-prevention caps cut out [to be used in Weebl]). Sitting on the box are +-400V/+580V power supply, filament transformer, input stage, and the new output stage. Dig the 3D (all over the place) component layout of the output stage!
Photo #2:
Focus on the tube/current source side of the output stage.
Transistors hanging in the air used to be mounted on the other side but that was only because their dissipation has been somewhat underestimated. The two without any heatsinks (pair of 2SA1967 supplying base current and setting offset for the anode current sources) could use a small stand-up or snap-on heatsink but do not seem to overheat without one either. Giant heatsinks for the anode current sources could safely be half this large as contrary to my expectations they only got mildly warm (~10W dissipation per each). 2SC3675 (pass transistor that creates -300V. ~6W dissipation) on two Aavid heatsinks joined together hanging nearby gets damn hot. But you see that large heatsink standing on the carpet? This transistor used to heat that one up just as well. Probably it just likes attention.
MOSFETs run warm but not hot. The ones driving the grid (2SJ79) are a little hotter than 2SK216s. They will get their small stand-ups eventually, but that'll be enough.
If crazy mounting (like this one) is used, the amp will oscillate in a very weird fashion and will refuse to listen to any adjustments. With Kevin's help, it has been mostly cured. Proper layout (due Christmas) should eliminate the remaining issues.
BTW, an easy way to see if any current is flowing through the tube is to look for the blue flame inside. Wicked stuff. Also, mono through Staxes (only 404 was used for testing) is not half-bad at all either.
