Yep, sorry I meant the small caps near the l7805.
(On another build using the sure amp, I was using a 24v battery pack to power the amp, and needed 5v to power the fan. I used nothing but a l7805cv in order to drop 24v down to 5v for a fan, and it all worked ok. I guess using a battery as a power source may have helped a little, regardless I may still add caps to this l7805 also).
Mike, be aware that 24V - 5V means 19 V drop on the poor 7805 ( do you use 78L05 or L7805 ? ). So let's assume such a small fan takes 50 mA at 5V ( check that please ). That means 19V x 0.05A = 0.95 W that is dissipated by the reg. That 78L05 will become hot continuously.... and its internal protection circuit will shut it down and when it has cooled enough it will power up again and the process will start over again. It will not live very long either when used like this.
We are way off topic here but for the sake of knowledge: if you want to drop such high voltages you better use the standard 7805 1 Amp version without the "L" in TO220 case and add a heatsink to it. Also you can use a resistor before the reg to drop some volts. Let's say 180 Ohm will drop 9 V on the resistor. Then the 7805 will have 15V - 5V = 10V drop. Then 10V x 0.05A = 0.5W is dissipated by the reg.... Possible with a standard TO220 7805.
The resistor will dissipate 9V x 0.05A = 0,45W. You better use a 1W resistor for that. You can play with the value of the resistor to have it dissipate more or less but always keep a minimum of 3 V drop over a 78xx reg otherwise it will not work correctly.
I hope you see that this is all a waste of energy to useless heat. And the purpose of the fan is to cool things down, right ? So you are overheating certain parts to cool other parts. It would be much wiser to find a fan that works on higher voltages with the same specs, rpm etc. I leave the math for ( for example ) a 12V fan to you. Just calculate and see how things will change. Components that stay relatively cool will also live longer. Just think of the 78L05 breaking down:no cooling anymore for the amp which will not like that etc. etc.
And of course the decoupling cap at the input should be near the reg not before the resistor ! The same counts for the output decoupling cap.
We are way off topic here but for the sake of knowledge: if you want to drop such high voltages you better use the standard 7805 1 Amp version without the "L" in TO220 case and add a heatsink to it. Also you can use a resistor before the reg to drop some volts. Let's say 180 Ohm will drop 9 V on the resistor. Then the 7805 will have 15V - 5V = 10V drop. Then 10V x 0.05A = 0.5W is dissipated by the reg.... Possible with a standard TO220 7805.
The resistor will dissipate 9V x 0.05A = 0,45W. You better use a 1W resistor for that. You can play with the value of the resistor to have it dissipate more or less but always keep a minimum of 3 V drop over a 78xx reg otherwise it will not work correctly.
I hope you see that this is all a waste of energy to useless heat. And the purpose of the fan is to cool things down, right ? So you are overheating certain parts to cool other parts. It would be much wiser to find a fan that works on higher voltages with the same specs, rpm etc. I leave the math for ( for example ) a 12V fan to you. Just calculate and see how things will change. Components that stay relatively cool will also live longer. Just think of the 78L05 breaking down:no cooling anymore for the amp which will not like that etc. etc.
And of course the decoupling cap at the input should be near the reg not before the resistor ! The same counts for the output decoupling cap.
Last edited:
Right now I do have the L7805 heatsinked to the metal chassis, but you are right it is getting rather hot. I think 60mm fan I'm using takes around 0.1A at 12v, so I imagine it should use less current at 5v, maybe half so 0.5A. 19 * 0.5 = 0.95W
I'll know the current draw for sure when I get home, but I'm pretty sure it was less than 100mA when powering the fan.
Thanks again, I think adding a resistor between the battery and 7805 is a nice solution as it takes some of the heat away from the 7805.
I now have a proper 24v fan somewhere which I haven't tested yet, but I think I may be able to run that straight from the battery without any other components.
Yes, the 24V version will be the best solution as long as it is quiet. I was editing my previous post to tell this but I passed the max. time that a post can be edited. But you got it right except for the calculation you wrote down...
Maybe you can try to use the 24V fan with a 7820 reg ( so 20V ) or even a 7818 to lower its rpm somewhat if it is too noisy.
Maybe you can try to use the 24V fan with a 7820 reg ( so 20V ) or even a 7818 to lower its rpm somewhat if it is too noisy.
Last edited:
Turn-on popping is more likely caused by large power supply caps than input caps.
-dr_vega
Perhaps when you start the whole thing up in unmuted state, not if you have an up and running power supply, then the input caps are the main influence on the pop....
I've found 4.7uF caps cause a pop, but I'm also running 0.05uF input caps for my tweeter channels and there is no pop whatsoever (thankfully for my tweeter!)... So it maybe worth trying a smaller value input cap if you want less of a pop?
I've managed to test the 24v fan and it is a little noisy at 24v. I ran it at 12v also and it was much quieter and was still fast enough to keep things cool, so I think I'm going to go with a 7812, and maybe a resistor in front of the 7812 to take some of the heat away from the 7812.
10k
I've done a lot of listening to passive volume controls at various input impedances and the lower the better down to the point where your source will start to have a hard time driving into it. I use 4k switched shunt attenuators which will be similar to a 10K pot.
Direct out into 4K
I use AK4395 dac chips direct out through foil caps with 4k stepped attenuators. 100K will work with a buffer or if there is no wire following it but 10k will sound more dynamic and have better extension at both frequency extremes.
Tried transformers
I tried Sowter transformers. They had a nice, easy, liquid migrange. The Dayton foil caps are more detailed and transparent and much cheaper for 6 channels in a DCX or M-Audio sound card.
Thanks a lot. I'll find 10k then
BTW Try out outputing the signal from AKM directly through transformers like LL1517 (even very, very cheap Monacor LT-110) . Amount of detail is insane !!
I tried Sowter transformers. They had a nice, easy, liquid migrange. The Dayton foil caps are more detailed and transparent and much cheaper for 6 channels in a DCX or M-Audio sound card.
Amp gain
I haven't tried it yet but it might be possible to reduce the gain of the amp to match the sensitivity of the drivers and reduce the hiss by changing the feedback resistors RFBC and using a lower voltage supply. This would eliminate the need for the big series resistor on your tweeter and give you back your sound. Many class AB amps have a minimum gain below which they become unstable, oscillate, and burn up. Usually around 22db of gain. I don't see any reference to a minimum gain for the TK2050 chip set and have posted the question but didn't get any reply. The only reference in the data sheets talks about matching the supply voltage to the RFBC so it would seem that one could take R13, 17, 25, and 31 all the way down to 7K ohms with a supply of less than 16v which will reduce the gain of the amp 6db, down to 18db, 8V/V taking the noise down with it and matching the drivers.
I haven't tried it yet but it might be possible to reduce the gain of the amp to match the sensitivity of the drivers and reduce the hiss by changing the feedback resistors RFBC and using a lower voltage supply. This would eliminate the need for the big series resistor on your tweeter and give you back your sound. Many class AB amps have a minimum gain below which they become unstable, oscillate, and burn up. Usually around 22db of gain. I don't see any reference to a minimum gain for the TK2050 chip set and have posted the question but didn't get any reply. The only reference in the data sheets talks about matching the supply voltage to the RFBC so it would seem that one could take R13, 17, 25, and 31 all the way down to 7K ohms with a supply of less than 16v which will reduce the gain of the amp 6db, down to 18db, 8V/V taking the noise down with it and matching the drivers.
Hi teamacc,
Metal film is not recommended by everyone in audio. Some state that metal films are too bright - while a good carbon film can sound very organic. It's just a matter of choosing the best combination for your amp and taste.
I'll have a look at removing d1 and d2 when I've finished building the amp. Do these need shorting or keeping as open circuit? My build is taking a while as spare time is limited - but it's coming together nicely.
Andrew
Hi, guys...
Just got one of those amp and started with mods
I had an idea with the PSU routing on the amp, so I need your assessment, is it necessary and will it work:
The idea is to separate PSU for each channel, more accurately to supply each channel with it's own rail... (better channel separation??) So, 4 wires from SMPS to the board, and an extra cap for each rail directly on the board connectors. Note that i won't be using 2 PSU's, only 2 rails from 1 PSU. Will this work?
I'm a noob, so please don't be hard on me
Just got one of those amp and started with mods
I had an idea with the PSU routing on the amp, so I need your assessment, is it necessary and will it work:
An externally hosted image should be here but it was not working when we last tested it.
The idea is to separate PSU for each channel, more accurately to supply each channel with it's own rail... (better channel separation??) So, 4 wires from SMPS to the board, and an extra cap for each rail directly on the board connectors. Note that i won't be using 2 PSU's, only 2 rails from 1 PSU. Will this work?
I'm a noob, so please don't be hard on me
I understand that it won't be the same as using two boards, I just wanted to know is it even worth it... It's no problem cutting 2 traces and linking the rest.
Also elcos were in plan, together with 0.1uF film cap in parallel under each tp2050.
Just fried one of my boards: one channel was really weak, i desoldered and resoldered the chip (tp2050, yes, the old one), and i think i made an error, as there were flames, smoke and smell.
Inspection revealed that 4 pins evaporated.
Luckily, it was the weak channel. I will try to solder the strong channel's chip back on tomorrow.
Inspection revealed that 4 pins evaporated.
Luckily, it was the weak channel. I will try to solder the strong channel's chip back on tomorrow.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.