E19 headphone amplifier board K2381 J407 MOSFET Yuanjing Audio

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I have a Sennheiser HD598 headphone currently used with the headphone amp section of a Tianyun Zero DAC with very good results. Just for the fun of it, I decide to build a solid state separate headphone amp. I found the YuanJing assembled board at a very good price.

https://www.yuan-jing.com/e19-class-a-headphone-amplifier

I was attracted to its all descrete, MOSFET output section design. It has uPC1237 Speaker Protection Chip. The producer indicated high quality parts are used. But I cannot find any schematics or circuit information.

I picked it over the Lehmann or Beyerdynamics clone. I already ordered one and it should be here in 2-3 weeks. Anyone has tried it before?

Edited September 15, 2016:

The E19 headphone amplifier board is NOT recommended for reasons that nattawa posted in #2. The hum problem is fixable, but you have a better, similarly priced, choice in Weiliang E5 board.

I wanted a front-end input using Marantz HDAM input structure,2SK170 matching of 1% for differential input, fully complementary MOSFET output, uPC1237 Protection Chip, and at least 500 mW into 50 ohms loads. The E5 board only is getting difficult to find, but fully assembled version with case is still available at good price.

http://www.ebay.com/itm/Finished-E5-Class-A-hifi-Headphone-amplifier-with-ALPS-potentiomer-115V-230V-/191869579640?hash=item2cac517178
 
Last edited:
I hope you did not get the same E19 that I ordered at ebay. It went on sale for too cheap to let pass so I took the bait, knowing there could be a "got'cha" moment. My son snatched The Wire from me and would not return it any time soon. I needed something quickly to drive those hungry 600-ohm K240DF cans. For $17 shipping included I was prepared to receive some fixer-upper.

So I wasn't much surprised by this constant buzzing noise when powering it up, but went on troubleshooting right away. The noise is a steady, stable, offensive, nasty sound that would not change by volume pot, or by touching sensitive spots on the board, or by the input signal being present/absent. It did surprise me when I scoped it and saw a 120Hz saw tooth-like wave form, about 18mV-pk, at the GND terminal of the phone jack!

A closer inspection quickly uncovered a missing capacitor that was the cause of the noise.

uPC1237 is supplied by a 12V auxiliary rail regulated by a 7812 3-terminal regulator, which is in turn supplied by a dedicated full-wave rectifier separate from that for the amplifier. However, a smoothing capacitor, supposedly following right after the rectifier diodes, was NOT there. So the rectifier sends 120Hz full wave pulse train straight to the 7812, and the charge current on the GND return line back-drives the headphones to make the noise -- the phone jack shares the same GND return path with the uPC1237 circuitry.

The remedy is simple and effective. I put in a 220uF/25V e-cap between the rectifier diodes and the centre tap of the AC entry, and the noise is gone for good. I had to install the cap on the bottom side of the board, as the designer missed it out entirely.
 
Unfortunately in certain parts of the world there is no concept of "prototype".
What in part of the world there is prototype-1, prototype-2, ..., release-v1, release-v2, in other parts prototype-1 is effectively release-v1, with customers having to suck it up.
 
My guess is they have failed that model of headphone amp and are dumping the batch of failed prototype builds to the online market, which in my book would be a malicious conduct.

Keilau, you'd better check the polarity of all the electrolytic caps on the board before powering it up. I've just had one of the four 220uF/50V electrolytic caps squirt on me. It was installed backwards.
 
Nattawa, thank you for 2 very helpful post. When I first got the board, the quality seemed to be good with clean soldering. I tried to sketch a schematic of the circuit first, but soon gave up because the 2-sided board has too many traces that cannot be seen.

I definitely will implement the mod you mentioned. It is so obvious once you pointed it out. You did a good job helping other DIYers.

I noticed that the hole spacing was for an ALPS RK16 volume pot, but the maker put a tiny one instead. It makes the volume control shaft way too low for normal mounting. I will start the assembly once the RK16 arrives.

I will use a 30VA toroidal transformer and a 2104 aluminum enclosure. I paid a little more than you did for the board. The total is still under $100 so far. I will post some pictures the next few days.
 
My guess is they have failed that model of headphone amp and are dumping the batch of failed prototype builds to the online market, which in my book would be a malicious conduct.

Keilau, you'd better check the polarity of all the electrolytic caps on the board before powering it up. I've just had one of the four 220uF/50V electrolytic caps squirt on me. It was installed backwards.

nattawa, thank you for the warning and I can use some help here.

The 4 big 6800uF caps are for the main power supply and seemed to be hookup consistently? For each rail, they connect 2 caps in SERIAL which effectively half the capacitance to 3400uF. If I remove one of the cap and jumper the trace, I can double the capacitance. Right? Very, very strange arrangement.

There are 6 other smaller caps nested among the heatsinks. I cannot tell which way the four 220uF/50V is supposed to go. Can you show a picture of which of them are reversed in polarity so that I can double check mine?

There are 4 trim pots. I supposed that they are for DC nulling and bias current. But there is no test point clearly marked. I will have to figure that out.

This cheap board turns out to be more fun that I bargain for.
 
volume pot replacement

I noticed that the hole spacing was for an ALPS RK16 volume pot, but the maker put a tiny RK09 instead. It makes the volume control shaft way too low for normal mounting. I will start the assembly once the RK16 arrives.

$_3.JPG
 
......

The 4 big 6800uF caps ...........

The schematic of the E19 is almost identical to that of E5 found HERE. Some component value may differ, and the output devices are a different type too.

The capacitor I blue-circled in the attached picture is the one (and only one) that got installed backwards.

I changed out the volume pot the first moment after I un-boxed the package. Not only was the original small, it was a linear taper B-type pot, wrong taper for the application, and it was installed crooked as well.

Big caps were removed for tracing out the PCB design. I had a huge curiosity on how they could have messed up the design:D.
 

Attachments

  • before.JPG
    before.JPG
    450.3 KB · Views: 1,232
The schematic of the E19 is almost identical to that of E5 found HERE. Some component value may differ, and the output devices are a different type too.

The capacitor I blue-circled in the attached picture is the one (and only one) that got installed backwards.

I changed out the volume pot the first moment after I un-boxed the package. Not only was the original small, it was a linear taper B-type pot, wrong taper for the application, and it was installed crooked as well.

Big caps were removed for tracing out the PCB design. I had a huge curiosity on how they could have messed up the design:D.

Yes, I did notice the similarity to the E5 too. I was drawn to the E19 by the larger power cap and lower price. Well, I ended up with 3400uF on the E19 vs. 9400uF for the E5. I am seriously considering removing the 4 6800uF and put 2 axial 10000uF/35V instead.

All the 220uF caps on my board were installed correctly. Thank you for the picture.

Did you play with the bias current adjustment?
 
Done tracing the power and signal returns, or the GND system. Turns out they did it the way a flash light wiring would be done in the power supply section, leaving signal returns share common copper trace impedance with power filter return lines at several places.

So I went ahead and cut many traces and modded the power supply into a C-R-C construction, having a clean "star-GND" where signal and power filter returns hook up straight to on their own wires. The amp is now dead quiet with volume pot at max and input open circuit on Senheisser HD600. I may have gone too far, and I don't have a lower impedance and more sensitive phones to verify or prove improvement. I may have to order another E19 amp just to make some A-B experiment.
 

Attachments

  • gnd.JPG
    gnd.JPG
    349.2 KB · Views: 1,175
  • after.JPG
    after.JPG
    396.7 KB · Views: 1,170
I set the output bias at about 60mA per transistor and be done with it

I wouldn't worry about the power filter capacitance. 4x 6,800uF is more than plentiful. In my C-R-C rail filter configuration that supplied with a 20VA 2x12Vac Antek toroidal, the 120Hz saw-tooth is about 120mV in amplitude on the first cap, and quarter that after 0.5-ohm on the second cap. I could go with a 10VA 2x 15Vac tranny for its much larger internal resistance if the ac hum was a challenge. For a Class A headphone amp where load regulation is not much of a concern, a smaller power rating transformer often widens the conducting angle and lowers the charge current peak, making a less noisy PSU.
 
My guess is they have failed that model of headphone amp and are dumping the batch of failed prototype builds to the online market, which in my book would be a malicious conduct.
Yes, exactly what I meant.
Many just grab either public schematics around, or reference ones out of manufacturer data sheets, which would work, if the devil within the detail would not play them tricks.
And the devil many times comes in the form of proper ground routing, or PCB signal routing in general, or properly designed power supply (which is mostly omitted by public or reference schematics), or/and choice of cheap/fake components.
 
Yes, exactly what I meant.
Many just grab either public schematics around, or reference ones out of manufacturer data sheets, which would work, if the devil within the detail would not play them tricks.
And the devil many times comes in the form of proper ground routing, or PCB signal routing in general, or properly designed power supply (which is mostly omitted by public or reference schematics), or/and choice of cheap/fake components.

That's why we need this forum here to sort out the pros and cons of many offers on Ebay or various websites. Nattawa did a wonderful job helping other users to make the E19 to work.

Among the FET input and full complimentary MOSFET output headphone amplifier boards, which is the one that meets your standard? Do you mind to share your experience too?
 
Last time I bought a "kit", it was many many years ago, wasn't on ebay/amazon/alibaba, and I don't even remember its specs.
Nowadays I either design them from the ground up (schematic, pcb, ... - for fun), or I buy directly brand ones.
The reality is, unless the design and/or implementation are really wrong, they will all have have flat-at-glass-tabletop frequency response within the audio bandwidth, and a distortion which is MUCH less than what the human ear can distinguish.
Given that, when I do design them, I like the challenge of achieving the "I can't tell the difference" result, with as much simplicity (in terms of design and number of components) as possible.

So, sorry but I can't suggest you any preferred ebay kit. Some sellers do follow up, and show up in forums like this one, but the reality is, out of a $30 kit, do you want to spend another $20 for shipping back to China (and wait at least a month of turnaround)?
 
HotIce, most of us participating here are hobbyists with various degree of understanding in electrical and electronic basics. Most are not well educated to do their own design. It may be true that yours is better than the rest of us. Find a way to contribute for the rest of mere mortals.

Nattawa showed how one can make lemonade out of a lemon, the E19.
 
HotIce, most of us participating here are hobbyists with various degree of understanding in electrical and electronic basics. Most are not well educated to do their own design. It may be true that yours is better than the rest of us. Find a way to contribute for the rest of mere mortals.
Where did I say, or even imply, that my design are better than "the rest of you"?
You asked for a suggestion for an ebay kit, and I told you I am sorry, but I prefer not to buy ebay kits, and hence have no experience to suggest one.
There are countless schematics on this site, with people sometimes offering PCBs.
If not, you can have PCBs made for cheap (using the posted Gerber files), and head to Digikey or Mouser to pick the various components.
Clearly, it is going to be more expensive than an ebay kit.
But if you consider the probability of getting a lemon by buying ebay kits, and you include that into the cost, that "more expensive" might suddenly look not so expensive after all.
 
Several different amps have been mentioned here.

Is the E5 as good or better than the E19?

Is the more expensive "Reference HA5000 amp" suppose to be better than the E5 and E19?

With all the hacking and reworking needed for the E19 board, I think I'm going to pass on it.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.