Ah ok so it acts as a heatsink 🙂 I was going to tap off the terminal blocks and use that source to drive fans blowing across mosfets and their heatsinks, maybe i can direct it in such way that they also target the jfets?
Thx again Salas !
Better avoid forced cooling. It can be uneven. Motors can send back EMI to the audio PSU. Use an independent 78XX chip for a fan if it will be mandatory.
What you use to insulate the devices from the heatsink is largely immaterial. The improvement comes from keeping the devices cool. OK mica works better as a thermal conductor than silicon rubber but to say that the mica makes for a better sound is very misleading and not entirely accurate.
If modulation of the device junction temperature can lead to poorer sound performance, then it follows that excessive and avoidable modulation of device temperature will result in a poorer sound performance than when a device is held at near constant temperature.
Whether Mica, or filled rubbers, or Kapton, or other/s will give excessive temperature modulation is up for argument.
That excessive temperature modulation can cause sound performance degradation is, in my view, not up for argument. I am not referring in this situation to reduced reliability affecting sound output.
Whether Mica, or filled rubbers, or Kapton, or other/s will give excessive temperature modulation is up for argument.
That excessive temperature modulation can cause sound performance degradation is, in my view, not up for argument. I am not referring in this situation to reduced reliability affecting sound output.
Exactly. It's got nothing to do with what you use as an insulator but all to do with keeping the devices at a lower stable temperature. Using big heatsinks with no insulator will keep everything cooler.
In the B1 everything is pretty stable. The B1 itself draws only a minimal load compared to the shunt regulation. There are not going to be any transient loads.
In the B1 everything is pretty stable. The B1 itself draws only a minimal load compared to the shunt regulation. There are not going to be any transient loads.
Last edited:
Next we will see that using steel bolts in an aluminium enclosure will cause noise due to differential metals.
BUT, using copper bolts to attach the devices to the heatsink might help get heat away from the device and into the heatsink.
BUT, using copper bolts to attach the devices to the heatsink might help get heat away from the device and into the heatsink.
Last edited:
Coupling capacitors
I think that I need some coupling caps on the outputs of my DCB1 because both my power amps have no coupling caps in the inputs. Just to be on the save side. Maybe I will make some DC coupled outputs and some without DC coupling in case I don't need them. I'll see.
this is what Salas wrote some time ago
My amps are both a 100K Zin amps. According to the Z-Cap website I can go down to a value as low as 0.80uF for my 100K Zin amps and have an optimal low frequency response at 20Hz. They even state that this is better than having higher value caps and thus having a lower optimal frequency response (Hz) and risking that some phase anomalies may introduced into the signal. Is this correct and how does this work? Just trying to understand... Lower values coupling are generally much cheaper, so please let it be correct 😀
I also read some things about using a resistor between the coupling caps and GND. So,
Do I need to put the caps as close to the DCB1 outputs as possible or is it OK to put them at the RCA chassis connectors?
Sorry of all these questions..
I think that I need some coupling caps on the outputs of my DCB1 because both my power amps have no coupling caps in the inputs. Just to be on the save side. Maybe I will make some DC coupled outputs and some without DC coupling in case I don't need them. I'll see.
this is what Salas wrote some time ago
My amps are both a 100K Zin amps. According to the Z-Cap website I can go down to a value as low as 0.80uF for my 100K Zin amps and have an optimal low frequency response at 20Hz. They even state that this is better than having higher value caps and thus having a lower optimal frequency response (Hz) and risking that some phase anomalies may introduced into the signal. Is this correct and how does this work? Just trying to understand... Lower values coupling are generally much cheaper, so please let it be correct 😀
I also read some things about using a resistor between the coupling caps and GND. So,
- where do I need to put these resistors, before or after the caps?
- Is there a way to calculate the right values for these resistors depending the the value of the caps or perhaps the input impedance of the power amps?
Do I need to put the caps as close to the DCB1 outputs as possible or is it OK to put them at the RCA chassis connectors?
Sorry of all these questions..
Check what happens with your power amp first.
How much does the output offset vary (or not) when different input resistances are connected.
Try zero ohms, try infinite ohms (open input), 220r (to mimic DCB1).
I have standardised on ~90ms for the high pass filter on all my power amps. It did not take long to swap the blocking capacitors to discover what I could not hear as any further improvement.
Your 100k for Zin and 0.8uF equates to 80ms, so is likely to be as low as you need to go.
BTW,
keeping the same capacitor and varying the Zin by adding a second 100k (to Signal Ground) to give 40ms High Pass then adding 51k for 20ms and then adding 24k for 10ms allows you to compare the effect of different input filters.
You could make the added resistor switchable.
How much does the output offset vary (or not) when different input resistances are connected.
Try zero ohms, try infinite ohms (open input), 220r (to mimic DCB1).
I have standardised on ~90ms for the high pass filter on all my power amps. It did not take long to swap the blocking capacitors to discover what I could not hear as any further improvement.
Your 100k for Zin and 0.8uF equates to 80ms, so is likely to be as low as you need to go.
BTW,
keeping the same capacitor and varying the Zin by adding a second 100k (to Signal Ground) to give 40ms High Pass then adding 51k for 20ms and then adding 24k for 10ms allows you to compare the effect of different input filters.
You could make the added resistor switchable.
1uF standard value capacitors work fine for coupling to 100K in practice IMHO. That is for plastic film. For electrolytics as couplers there are bass region THD issues due to their losses with rising signal level that is why you may spot up to even 220uF or 470uF in some professional audio preamps. Especially when they have EQ options like in studio mixing desks.
I can try varying with different input resistances to see how both my amps behave on the outputs, but I'm not entirely sure what I should do with the measurements. Please explain.
I was also thinking of using 1uF caps, Mundorf Evo Oil caps to be exact. My diy headphone amplifier has some Nichicon muse ES 470uF electrolytics in the signal path. They are bypassed by some 0.22Uf Spraque Vitamin Q's. This is a good sounding combination. I still have some of the Nichicon's left so I can compare them to fi the Mundorf's when I decide to use coupling caps.
If your output offset is near zero mVdc and does not vary with changes in Rs then that shows your power amp is not susceptible to variations in your source equipment output impedance. Do nothing.
If the output offset varies then you must add a DC blocking cap to the Output of the Source equipment, or to the Input of the Power amplifier. Do not add two DC blocking capacitors, each channel only needs one. It does not matter if the DC blocking cap is in the Source, or in the Receiver.
I prefer to add this blocker to the Input of the Receiver. That way I know I can connect any Source and not risk blowing up my speaker.
If the output offset varies then you must add a DC blocking cap to the Output of the Source equipment, or to the Input of the Power amplifier. Do not add two DC blocking capacitors, each channel only needs one. It does not matter if the DC blocking cap is in the Source, or in the Receiver.
I prefer to add this blocker to the Input of the Receiver. That way I know I can connect any Source and not risk blowing up my speaker.
How about the motorcycle stuff?? 😀
BK
An externally hosted image should be here but it was not working when we last tested it.
Motorcycle lonely trips are a poetic experience so the veteran Ducati battery is going to bless the Vref LEDs with Italian finesse noise for ethereal highs. 🙂
"Under neon loneliness motorcycle emptiness"
"Under neon loneliness motorcycle emptiness"
Salas, you are a genius and I believe that you are correct. Look at the last three digits of the DMM, which happen to spell out the Ducati model the battery came from. Spooky and blessed indeed. No other explanation!
BK
In most commercial equipment you will find that the output of one line amp will have a cap and the input to the power amp will also have a cap.
There is no harm in having more than one cap but it is unnecessary.
There is no "harm" in that it will not damage your equipment.
But !!!!
if your Receiver is Zin=100k and Cin=1uF then this combination has an F-3dB ~ 1.6Hz
If you use a DC blocking cap at the output of the source as well as that 1uF in the Receiver the effective capacitance is the series pair of caps. Let's assign an Rs=200r and Cout=470nF.
The F-3dB is now set by the 100k & 329nF. The F-3dB has moved to ~5Hz.
One has chosen a sensible input capacitor for the Zin of the receiver and the Source had a not uncommon output combination.
Let's now change the YOUR receiver and stay with the same Source.
Set Zin=22k and Cin=4u7F the F-3dB is still around 1.5Hz
But this time using the second DC blocking cap in the unchanged Source the effective capacitance is 420nF. The F-3dB has moved to 17Hz. This is more than a decade higher than what was intended when you chose the High Pass filter for your Receiver.
If both Source and Receiver have DC blocking capacitors, then in my view it is sensible to ADD bypass socket where one or other DC blocking capacitor is removed from the interconnection. This allows the user to choose the "better" high pass filter capacitor for the system.
But !!!!
if your Receiver is Zin=100k and Cin=1uF then this combination has an F-3dB ~ 1.6Hz
If you use a DC blocking cap at the output of the source as well as that 1uF in the Receiver the effective capacitance is the series pair of caps. Let's assign an Rs=200r and Cout=470nF.
The F-3dB is now set by the 100k & 329nF. The F-3dB has moved to ~5Hz.
One has chosen a sensible input capacitor for the Zin of the receiver and the Source had a not uncommon output combination.
Let's now change the YOUR receiver and stay with the same Source.
Set Zin=22k and Cin=4u7F the F-3dB is still around 1.5Hz
But this time using the second DC blocking cap in the unchanged Source the effective capacitance is 420nF. The F-3dB has moved to 17Hz. This is more than a decade higher than what was intended when you chose the High Pass filter for your Receiver.
If both Source and Receiver have DC blocking capacitors, then in my view it is sensible to ADD bypass socket where one or other DC blocking capacitor is removed from the interconnection. This allows the user to choose the "better" high pass filter capacitor for the system.
These showed up today. Yummy.
BK
BK
An externally hosted image should be here but it was not working when we last tested it.
Nearly there.
BK
BK
An externally hosted image should be here but it was not working when we last tested it.
If you will like it as it is with the Auricaps and they will stay, trim their leads shorter in the end. Nice work in progress photos.
Hi I made mine without plastic caps. I allmost never use Caps starting with M . But my version is very very booring. So maybee I should try in this construction.