Re: Re: GB150 D5, D6.... TO92 be gone.
All is going well and during setting it stabilized faster and once the bias was set did not drift anywhere as much as previously. The temperature is certainly no warmer than before and we've had some hot days.
I did have a problem with 1 of my modules, burning the safety resistor a few times, prior to this change and Greg thought it could have been thermal run away. He even went to the trouble of testing this module for me and found no problems with it. This spurred on the change and the desire to have a solid mechanical mounting. The problem has not returned and all is cool.
Primetime said:
All is going well and during setting it stabilized faster and once the bias was set did not drift anywhere as much as previously. The temperature is certainly no warmer than before and we've had some hot days.
I did have a problem with 1 of my modules, burning the safety resistor a few times, prior to this change and Greg thought it could have been thermal run away. He even went to the trouble of testing this module for me and found no problems with it. This spurred on the change and the desire to have a solid mechanical mounting. The problem has not returned and all is cool.
Thanks Rabbitz & MX-5,
Looks like BD139 & BD140 is the way to go. Did you guys tried tweaking with the premium stuff like BG/MUSE for the feedback caps, low noise resistors, silver mica caps, or increase the value for C4/5. Any audible difference in the sound? Worth the extra?
I bought the GB150D & GB300D in one go.
... I bet Greg must be very happy with this decision..... But he better make this worthwhile.. 
Frankly, 300W is quite lot (for my neighbour
)... MX-5, are you running yours full +/- 75V? I'm still deciding which transformer to buy. Just wondering if there is any advantage running the GB300D at a lower power.
Looks like BD139 & BD140 is the way to go. Did you guys tried tweaking with the premium stuff like BG/MUSE for the feedback caps, low noise resistors, silver mica caps, or increase the value for C4/5. Any audible difference in the sound? Worth the extra?
I bought the GB150D & GB300D in one go.
... I bet Greg must be very happy with this decision..... But he better make this worthwhile.. Frankly, 300W is quite lot (for my neighbour
)... MX-5, are you running yours full +/- 75V? I'm still deciding which transformer to buy. Just wondering if there is any advantage running the GB300D at a lower power.yes.
the 300 is able to run to a maximum power of 300W into the specified load.
With a high voltage supply the minimum load impedance will be 8ohm (Greg clearly states this). If one chooses a lower voltage then lower impedances can be used.
6ohm or 4 to 8ohm would be next and 4ohm capable from the lowest supply voltage option.
If the 300W into 4ohm option is chosen then the amps will still manage just over 150W into 8ohm and deliver good sound into the higher impedance. (you'll gather I'm not a fan of low impedance loads).
Re: GB150 D5, D6.... TO92 be gone.
rabbitz, mx-5
I took some measurements earlier today:
The transistors are mounted in ~4.8mm dia. holes, 4mm deep.
Room temp was 21degC
Notes:
Voltage across 10R drifts 0.13v over 25 mins without input and load.
With input and load attached the average voltage increases by 0.33v (from 0.66 to 1v). Voltage drifts 0.14v over 30 mins. I have ignored first two interval measurements where bias has not stablized.
However since the drift figures above are only calculated using the figures measured at the specified intervals, it doesn't show the real drift neccessarily. i.e. after running the amp a little harder to raise heatsink temp to 40degC I used the multimeter's min/max feature for a few minutes which showed the real drift (the DMM measures in 100ms intervals) as 0.27v! (min was 0.84, max was 1.11v).
Greg has states that this 0.14v (14%) is what he would expect for reasonable coupling after warmup.
Comments appreciated, I'm not sure what to do - and I know there are others interested hence I posted the measurments.
rabbitz, mx-5
I took some measurements earlier today:
Code:
w/o input and load with input/load
Minutes Voltage Temp Minutes Voltage Temp
0 0.60 22.0 0 0.67 28.9
1 0.67 22.0 5 0.85 35.0
2 0.68 22.2 10 1.05 36.6
3 0.63 22.2 15 0.98 37.3
4 0.73 22.9 20 1.05 38.3
5 0.72 23.3 30 0.91 39.6
6 0.64 23.6
7 0.62 23.9
8 0.64 24.2
9 0.64 24.5
10 0.67 24.8
15 0.69 26.3
20 0.68 27.4
25 0.69 28.3
avg 0.66 24.1 avg 1.00 37.95
min 0.60 22.0 min 0.91 36.6
max 0.73 28.3 max 1.05 39.6
diff 0.13 6.3 diff 0.14 3.0The transistors are mounted in ~4.8mm dia. holes, 4mm deep.
Room temp was 21degC
Notes:
Voltage across 10R drifts 0.13v over 25 mins without input and load.
With input and load attached the average voltage increases by 0.33v (from 0.66 to 1v). Voltage drifts 0.14v over 30 mins. I have ignored first two interval measurements where bias has not stablized.
However since the drift figures above are only calculated using the figures measured at the specified intervals, it doesn't show the real drift neccessarily. i.e. after running the amp a little harder to raise heatsink temp to 40degC I used the multimeter's min/max feature for a few minutes which showed the real drift (the DMM measures in 100ms intervals) as 0.27v! (min was 0.84, max was 1.11v).
Greg has states that this 0.14v (14%) is what he would expect for reasonable coupling after warmup.
Comments appreciated, I'm not sure what to do - and I know there are others interested hence I posted the measurments.
Hi Vikash
My GB150 runs with the bias set at 1V over the 10R resistor so would place higher demands on the heatsink. I ended up with an increase of 22°C after pushing the amp for a few hours (28°C ambient when testing). Both modules are mounted on the same heatsink which is 0.37°C/W rated.
With a lower bias, non Aussie climate and larger heatsinks which I think you have, yours is going to run cooler than mine. IMO, I can't see any hassles with yours but I'm no expert. If Greg says it's cool (no pun intended), then all should be well.
I just had to change mine to a different mounting due to my idiosyncrasies. OCD tends to make people worry about the most stupid things.
To summarise.... set, forget and throw the lid on.
My GB150 runs with the bias set at 1V over the 10R resistor so would place higher demands on the heatsink. I ended up with an increase of 22°C after pushing the amp for a few hours (28°C ambient when testing). Both modules are mounted on the same heatsink which is 0.37°C/W rated.
With a lower bias, non Aussie climate and larger heatsinks which I think you have, yours is going to run cooler than mine. IMO, I can't see any hassles with yours but I'm no expert. If Greg says it's cool (no pun intended), then all should be well.
I just had to change mine to a different mounting due to my idiosyncrasies. OCD tends to make people worry about the most stupid things.
To summarise.... set, forget and throw the lid on.
Hi Andrew. That is 60mA measured across the 10R resistor w/o input/load on first start up.AndrewT said:
After playing music (sine wave measuring 2.5V across 10R resistor) into 4R load for 10 mins I set the bias to ~100mA/1V (VR1 approx. centred). V across 10R fluctates by up to 0.35V within the space of (what feels like) a minute or so.
The 'at switch on' V with no load/input is now about 0.85 IIRC.
I've now done the zener mod, increased bias to 150mA, and installed the BD139/140. There's a definte improvement, but it still hunts a lot more than rabbitz's measurements (up to 0.15V), but at least it's halved from using BC546/556. I'm happy to move on now
When you guys measure hunt, over how many seconds/minutes do you take the readings?
V
When you guys measure hunt, over how many seconds/minutes do you take the readings?
V
More on hunting
-----------
Firstly, and like many, I'm learning as I go along so this thermal feedback, hunting, and how and when to measure it is all new to me and not as straight forward as I initially read
What I have deduced by trial and error today is that I can get consistent readings by first running the amp for about 5 minutes (I use sine wave so that voltage across 10R resistor measures ~3V for 5 mins). If I then pause the music the reading across 10R resistor gradually drops from ~2.5V to 1V (the bias I have previously set with VR1) over the next ~3-5 minutes. From this point on the measurement stays within +/- 0.06V. (I've only checked this over a couple of minutes).
So basically you can set bias only after reaching this point.
V
Good instruction, except the problem was not knowing when to actually set it as it kept drifting by so much.rabbitz said:
-----------
Firstly, and like many, I'm learning as I go along so this thermal feedback, hunting, and how and when to measure it is all new to me and not as straight forward as I initially read
What I have deduced by trial and error today is that I can get consistent readings by first running the amp for about 5 minutes (I use sine wave so that voltage across 10R resistor measures ~3V for 5 mins). If I then pause the music the reading across 10R resistor gradually drops from ~2.5V to 1V (the bias I have previously set with VR1) over the next ~3-5 minutes. From this point on the measurement stays within +/- 0.06V. (I've only checked this over a couple of minutes).
So basically you can set bias only after reaching this point.
V
I've set mine at 1V nomimal after a good warm up (time as per the instructions) and the amp has stabilised. The DMM will read values above and below 1V but I set on the average. It seems to come back to this point even after drifting up and down.
On your graph, mine would be set in the first stabilsed period after stopping the music. The dropping of bias V after that is irrelevant IMO as it's lower than the initial value so will not cause problems and soon as you play tunes, it's going to go back up anyway.
Can't remember the exact numbers now but would be set at say 0.97V min & 1.03V max. Occasionally you'll get a larger variation due to the mains supply or an air conditioning unit kicking in or out, but you just have to be patient and methodical.
I think Greg has quite a tolerance on the final bias and from my understanding can be set anywhere between 0.7V - 1.3V for good operation.
On your graph, mine would be set in the first stabilsed period after stopping the music. The dropping of bias V after that is irrelevant IMO as it's lower than the initial value so will not cause problems and soon as you play tunes, it's going to go back up anyway.
Can't remember the exact numbers now but would be set at say 0.97V min & 1.03V max. Occasionally you'll get a larger variation due to the mains supply or an air conditioning unit kicking in or out, but you just have to be patient and methodical.
I think Greg has quite a tolerance on the final bias and from my understanding can be set anywhere between 0.7V - 1.3V for good operation.
rabbitz, cheers for that. Greg also mentioned how delicate the sensing was to the surounding air flow so I measured again - but this time left it covered with a cardboard box and went off to lunch.
Sorted! Paul, are you reading?
V
An externally hosted image should be here but it was not working when we last tested it.
Sorted! Paul, are you reading?
V
input transistor emitter resistors
I had replaced all of the resistors in the signal path in my SKA with better quality ones - mostly Caddock TF020's to very good effect.
However recently when I had a fault on one of the channels I had another look at the diagram and noticed that the emitter resistors on Q1-4 had been overlooked by me and they were still cheap metal oxide types and in the feedback loop.
After some work in Spice I concluded that bypassing them with a wire should be OK ( I did not have any 22ohm high quality resistors ) .
Shorting out these resistors had two effects on the sound:
1) Much cleaner - because these were the last cheap resistors removed from the signal path throughout my system - this was expected
2) Much more dynamic ! - more than just the extra 3-4 db of open loop gain could explain I think. The amp now sounds much more exiting and rhythmic - this was unexpected.
Thought I would pass this on to all.
I would interested to hear if others notice the same dynamic effect. I would not expect you to notice the cleaner sound unless you have the same circumstances as me but I would strongly recommend using better quality resistors in all key positions to hear this also.
cheers
mike
I had replaced all of the resistors in the signal path in my SKA with better quality ones - mostly Caddock TF020's to very good effect.
However recently when I had a fault on one of the channels I had another look at the diagram and noticed that the emitter resistors on Q1-4 had been overlooked by me and they were still cheap metal oxide types and in the feedback loop.
After some work in Spice I concluded that bypassing them with a wire should be OK ( I did not have any 22ohm high quality resistors ) .
Shorting out these resistors had two effects on the sound:
1) Much cleaner - because these were the last cheap resistors removed from the signal path throughout my system - this was expected
2) Much more dynamic ! - more than just the extra 3-4 db of open loop gain could explain I think. The amp now sounds much more exiting and rhythmic - this was unexpected.
Thought I would pass this on to all.
I would interested to hear if others notice the same dynamic effect. I would not expect you to notice the cleaner sound unless you have the same circumstances as me but I would strongly recommend using better quality resistors in all key positions to hear this also.
cheers
mike
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.