Hello,
here some fotos of my pimping/tuning/tweaking of the cheap headphone amp Millenium HA4 from Thomann, it costs only 30 dollars/26 Euros. This item works just well by itself. No one tried to tweak it until now?? However I wanted to see what type of tuning can make it better.
Its fairly easy to lift the entering diode on one side in order to insert into the plus line one or two regulators with flying wires. You can glue the regulators into the side of the Millenium as there is very little space in it but it works. The metal chassis is not conducting, so there was no need of special isolation of the regulators. You can glue them directly to the metal. The regulators do not produce any heat as the current is low.
I took an existing 24 V source (original is 12 Volts) and have put a LM7818 and LM7812 together I soldered them directly to each other with their legs together, only the first leg has to be shortened in length for leaving open input and output for the wiring. I like the idea to get a better PSRR of over 100db, for this I cascaded the regulators but certainly a single one 7812 will do the job. The regulators became 0,22 mfd film caps between their legs for better ripple rejection. I also added the protecting diode from input to output for the regulators.
I also dropped in some different coupling capacitors like WIMA MKS 10mfd, 50 Volts, a 16 Volt 470mfd and a Panasonic FC 22mfd, 25 Volts. The coupling capacitor is on the summed negative line of each output to headphone. Original electrolytic is 16 Volt 100mfd.
So listening showed that the difference between the electrolytic capacitors is fairly small if not impossible to hear a difference. Only the foil type makes an audible difference (Wima 10mfd MKS). So listening with an earphone can clearly give you the chance of testing the audibility of different capacitors. What helps the sound a bit is putting a small foil on the Elko on the backside of the PCB. This is a lightweight solution without the need to lift the original 100mfd 16V capacitors what is a bit tricky as the desoldering is not quite easy.
Just do not forget to wire also the legs in the middle of the regulators to the minus line of the 1000mfd cap, I forgot that and added it later, the regulators were not harmed. So you need three wires, two for the plus line, one for the negative line.
So what does it do? Transient response is better and the augmented ripple rejection is clearly audible. The Millenium makes now much more fun. With my favourite Headphone AKG K240DF the sound is great for small bucks. Alternatively I listeh with the Thomann Superlux HD668B - big bang for the bucks, as the AKG is out of production.
I pimped two Millenium HA4, depending on the voltage source I have - I used for the other one a 15 Volt LM2940 (needs 22mfd in the output for stability) and good old 7812 in series. Interesting, you can clearly hear the difference in the PSU, one is a traditional 1A 15 Volt transformer-type (output 20,4 Volt), the other one a 24 volt, 1A, switching type. The ladder sounds more dry in the bass. I like the transformer-one better - more gentle.
Maybe this tweak is worth a try for someone.
picload.org | milleniumha4thomann3.jpg
picload.org | milleniumha4thomann1_.jpg
picload.org | milleniumha4thomann2.jpg
Dragan from Germany
here some fotos of my pimping/tuning/tweaking of the cheap headphone amp Millenium HA4 from Thomann, it costs only 30 dollars/26 Euros. This item works just well by itself. No one tried to tweak it until now?? However I wanted to see what type of tuning can make it better.
Its fairly easy to lift the entering diode on one side in order to insert into the plus line one or two regulators with flying wires. You can glue the regulators into the side of the Millenium as there is very little space in it but it works. The metal chassis is not conducting, so there was no need of special isolation of the regulators. You can glue them directly to the metal. The regulators do not produce any heat as the current is low.
I took an existing 24 V source (original is 12 Volts) and have put a LM7818 and LM7812 together I soldered them directly to each other with their legs together, only the first leg has to be shortened in length for leaving open input and output for the wiring. I like the idea to get a better PSRR of over 100db, for this I cascaded the regulators but certainly a single one 7812 will do the job. The regulators became 0,22 mfd film caps between their legs for better ripple rejection. I also added the protecting diode from input to output for the regulators.
I also dropped in some different coupling capacitors like WIMA MKS 10mfd, 50 Volts, a 16 Volt 470mfd and a Panasonic FC 22mfd, 25 Volts. The coupling capacitor is on the summed negative line of each output to headphone. Original electrolytic is 16 Volt 100mfd.
So listening showed that the difference between the electrolytic capacitors is fairly small if not impossible to hear a difference. Only the foil type makes an audible difference (Wima 10mfd MKS). So listening with an earphone can clearly give you the chance of testing the audibility of different capacitors. What helps the sound a bit is putting a small foil on the Elko on the backside of the PCB. This is a lightweight solution without the need to lift the original 100mfd 16V capacitors what is a bit tricky as the desoldering is not quite easy.
Just do not forget to wire also the legs in the middle of the regulators to the minus line of the 1000mfd cap, I forgot that and added it later, the regulators were not harmed. So you need three wires, two for the plus line, one for the negative line.
So what does it do? Transient response is better and the augmented ripple rejection is clearly audible. The Millenium makes now much more fun. With my favourite Headphone AKG K240DF the sound is great for small bucks. Alternatively I listeh with the Thomann Superlux HD668B - big bang for the bucks, as the AKG is out of production.
I pimped two Millenium HA4, depending on the voltage source I have - I used for the other one a 15 Volt LM2940 (needs 22mfd in the output for stability) and good old 7812 in series. Interesting, you can clearly hear the difference in the PSU, one is a traditional 1A 15 Volt transformer-type (output 20,4 Volt), the other one a 24 volt, 1A, switching type. The ladder sounds more dry in the bass. I like the transformer-one better - more gentle.
Maybe this tweak is worth a try for someone.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
picload.org | milleniumha4thomann3.jpg
picload.org | milleniumha4thomann1_.jpg
picload.org | milleniumha4thomann2.jpg
Dragan from Germany
1. It is advised to attach pictures here so that they do not get lost. --> "Manage Attachments"
If in doubt, the greater voltage headroom afforded by a single 7818 may be more useful.
BTW, I'd preferably use a 78M series part (500 mA max) for generally lower noise and better regulation at low currents... a 78L would probably be pushing it under load.
bekommen --> get, receive, ...
become --> werden!!!
Really common trap for German speakers, and well-documented as such.
BTW, I wouldn't be surprised if the 7818 were dropping out of regulation all the time when using the nominal 15V supply, assuming the 20.4 V was measured unloaded. A traditional regulator like that has a dropout voltage of about 1.8 V minimum, more like 2.0-2.1 V typical. So a 7818 needs about 20.0 V in. Now unregulated DC supplies have a habit of giving relatively high output unloaded but quickly going down under even a small load before "settling in" - and you only had 20.4 V to begin with. See the problem?
I'd try the same again with a 15 V regulator instead.
Getting >100 dB of PSR tends to require a good layout, not sure that's necessarily the case here...
If in doubt, the greater voltage headroom afforded by a single 7818 may be more useful.
BTW, I'd preferably use a 78M series part (500 mA max) for generally lower noise and better regulation at low currents... a 78L would probably be pushing it under load.
No, they received some.
bekommen --> get, receive, ...
become --> werden!!!
Really common trap for German speakers, and well-documented as such.
Yeah, because its capacitance is too small, causing a dropoff in L+R bass even with the 600 ohm K240DF (~equivalent of 5 µF per channel). The 470µ/16V would be my pick here... OK for 32 ohm loads even if amplifier output series resistance were zero, and I don't think it is, so just fine then. This single-cap design is not ideal, but makes for minimum parts count.
Likewise, you could add some more electrolytics on the back, too.
That's odd. Common-mode RF artifacts or capacitive mains coupling?
BTW, I wouldn't be surprised if the 7818 were dropping out of regulation all the time when using the nominal 15V supply, assuming the 20.4 V was measured unloaded. A traditional regulator like that has a dropout voltage of about 1.8 V minimum, more like 2.0-2.1 V typical. So a 7818 needs about 20.0 V in. Now unregulated DC supplies have a habit of giving relatively high output unloaded but quickly going down under even a small load before "settling in" - and you only had 20.4 V to begin with. See the problem?
I'd try the same again with a 15 V regulator instead.
Last edited:
Hello,
thank you for your reply!
Yes, you can hear that the 10mfd MKS could be a bigger one but it was the cap I had at hand. As I can choose the output of preference I was keen to take the risk ;-)
The 7818+7812 tandem has a 24 Volt source. The Lm2940,15V+7812 tandem has a 15 Volt transformer as a source and yes, 20,4 Volt was unloaded.
So maybe some better quality bipolar capacitor would make a more obvious change but they are not so easy to find and to make an order only for purchasing one special type of capacitor was not economic.
Bye, Dragan
thank you for your reply!
Yes, you can hear that the 10mfd MKS could be a bigger one but it was the cap I had at hand. As I can choose the output of preference I was keen to take the risk ;-)
The 7818+7812 tandem has a 24 Volt source. The Lm2940,15V+7812 tandem has a 15 Volt transformer as a source and yes, 20,4 Volt was unloaded.
So maybe some better quality bipolar capacitor would make a more obvious change but they are not so easy to find and to make an order only for purchasing one special type of capacitor was not economic.
Bye, Dragan
Hello,
here I found some information on cascading LDOs:
https://www.maximintegrated.com/en/app-n...mvp/id/883
They tell as a manufacturer that the specs are doubling which sounds great and is more I expected.
Greetings, Dragan from Germany
Citation:
Conclusion
An external network added to the input of a linear regulator improves the inherent PSRR of the LDO, especially at high frequencies, where the low quiescent current compromises the high-frequency PSRR of the LDO.
Of the methods discussed, the additional LDO is the most general-purpose, packs more attenuation into a given small area, and requires the least design time.
For low-current applications requiring only modest protection, the RC filter method is cost-competitive, but requires careful trade-offs in component selection.
For high-current LDO controllers, the addition of one resistor to the controller input supply very effectively enhances the PSRR of the entire circuit.
here I found some information on cascading LDOs:
https://www.maximintegrated.com/en/app-n...mvp/id/883
They tell as a manufacturer that the specs are doubling which sounds great and is more I expected.
Greetings, Dragan from Germany
Citation:
Conclusion
An external network added to the input of a linear regulator improves the inherent PSRR of the LDO, especially at high frequencies, where the low quiescent current compromises the high-frequency PSRR of the LDO.
Of the methods discussed, the additional LDO is the most general-purpose, packs more attenuation into a given small area, and requires the least design time.
For low-current applications requiring only modest protection, the RC filter method is cost-competitive, but requires careful trade-offs in component selection.
For high-current LDO controllers, the addition of one resistor to the controller input supply very effectively enhances the PSRR of the entire circuit.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.