Sunday, November 26, 2006

An Afternoon with the Tandberg 3001 Supply

Rather than try a bunch of things and have then be wrong, I ordered lots of spare transistors of the one's I could order, and I spent most of the afternoon today looking for the bad one in the 25V supply.

After replacing a host of little transistors I finally found the one that was causing the problems. The supply is 95% done. The critical 15V and 25V supplies are working perfectly. It wasn't a complete waste of time though, a lot of the transistors were being upgraded to 65V parts, which will make this unit much more stable should I ever have to remove it and modify it on the bench again.

Unfortunately, I'm still having serious problems with the two 5V supplies though. Let me share with you my current (haha!) problems. Take a look at the relevant part of the schematic:

To the left of the 33 Ohm/4W resistor is 15 Volts, as it should be. The problem I'm having is that the two transistors you see here should be outputting 5 and 5.2 V respectively. However, there is too much of a voltage drop through R822, leaving me with something like 2.8 V instead of the 5.9 I should expect. The problem seems to be the current draw through Q814(Q813 on the board). The way I see it, in order for the power supply to behave as documented, there should be about 4.1 mA through R827, and about 1.4mA through Q814. Instead, I'm getting 4.8 mA through Q814 alone. What I don't understand is WHY. The parts all seem to be working correctly. If I remove Q813, and replace Q814 with a brand new part, I still have the same problem. I can't get the voltage at the base of Q813 up to 5.9V. And from what I can see, besides actually changing the value of R827 I really don't have a chance. The input voltage remains a rock solid 15.04 volts throughout though.

I've removed the electrolytic, and measured all of the resistors. Everything is measuring what it should.

Suggestions? Is the schematic bogus to start with?

Modding The Rega EAR Headphone Amplifier

Even though I don't think the Rega EAR is worth the purchase new, I decided to put in information for those of you who may already own a Rega EAR and are kind of curious about getting into making your own modifications. This is actually a device that will be simple to mod for a few bucks which can actually yield some serious benefits.

Tools you'll need:

  1. Metric Allen wrench set
  2. Phillips screwdriver
  3. Soldering Iron
  4. Solder Wick
  5. Solder
You'll need the metric Allen wrench set to remove the front cover. You'll just strip them otherwise, they are on tight.

Take the back plate off. Before you do anything else, measure the space between the top of the board and the case. You'll need this to make sure you don't buy caps which are too tall.

If you read instructions on the internet and follow them without accepting responsibility for your actions, or without really knowing what you are doing and hurt yourself, burn down your house, ruin your stereo or go stark raving mad and start attacking your neighbors lawn mower with an ax while naked in the middle of winter, you are a certifiable dufus and should read no further.

If on the other hand you realise that you are taking risks for the functionality of the equipment you are modifying, and your own health and safety, and will take full responsibility for the consequences of your actions and you are not going to try to act like a dufus by blaming the author of this or any other article online, then please continue.

First, a note about how capacitors are measured. There are three mesuarements which are usually important. Capacitance, Voltage and Temperature. Capacitance is the only value which affects how the part works in circuit. Voltage and Temperature affect when the part will fail. If you put a cap in with too small a voltage rating, it will break down very quickly. For this reason, you can increase the voltage and temperature grades on a capacitor relatively safely. On the other hand, going too far over in voltage may cause you to use a part too physically large, which will be more inductive and may not physically fit. For instance, it's ok to replace a 35V Cap with a 50V or 100V cap. There are cases when the manufacturers original parts were rated too low (see my posts on the Tandberg 3001 power supply) in which case a voltage upgrade makes a really good idea. I don't think you'll find this a problem with the Rega.

One thing to be careful of is that electrolytic capacitors have a direction they should be used in or they will fail fast. Observe how they are mounted in the original, look for (+) signs on the silk screen. Make sure the (-) mark on the capacitor goes to the opposite pin.

Take every electrolytic you see, and replace them with Panasonic FM capacitors, available from Digikey. You should leave the values the same except for the the power supply and output caps. The power supply filter caps are the two large caps rated 1,000uF right next to the power supply jack. Buy the largest cap you can fit into their place. If you can fig 1,800uF you should. This isn't just upgrade-itis. This unit has some serious hum, the only way to remove it is going to be to increase the filter capacitance. By putting in low inductance caps, we can also hope to reduce some of the hiss as well.

For the output caps, putting in bigger units will give you better bass, at the expense of increased inductance and less treble. I suggest that since the Pana FM caps are so low in inductance anyway you should upgrade these by 10-20% if they will fit. You should be able to fit a 0.1uF metal poly film cap underneath each output capacitor. Check the case to board spacing first though.

For all other electrolytic caps, match the original uF values. Leave the ceramic and box caps alone.

After adding significant capacitance to the power supply you may blow the little fuse near the jack. Ce est le vie. Short the puppy and move on. Ok, no, just kidding. If this problem arises I would seriously do is add a fuse jack to the back of the unit, and wire it to where this thing is. Then try to put the smallest fuse which doesn't fail when you power the unit on. Somewhere around 200mA to 315mA 5x20mm fuse should work.

You can check the on board fuse for failure by testing the AC voltage across it. It should be near 0. If it's the same as the voltage coming from the wall wart, it's blown.

The single chip in the middle of the board towards the back is an NE5532 dual operational amplifier. Rega probably picked this particular op amp because it makes the rest of the circuit incredibly easy. Let me explain. For an op amp, the 5532 has very low output impedance and is able to drive 600 Ohms to +- 10 V easily. This means they can buffer the input and drive the current gain stage directly, without any more parts. In other words, it's the cheapest possible solution they could find and still have a large current gain capability.

Remove it and replace it with a high quality machined IC socket. This single upgrade is going to give you the most amount of fun, because you'll then be able to experiment with a variety of different dual op amps until you find the one that you like best without having to pull out the solder gun each time.

I suggest the first good one to try is an NE5535. It's probably got the closest sound of the NE5532, but with smoother, richer detail, while still being a rather "cold" part. Philips released it as a Class-A biased version of the NE5532, but I don't think they even bothered to use it in their own products very much and it consumed higher power so except for the modifications market it didn't make much of a dent in the market. I believe Walt Jung (Walt, e-mail me if I'm wrong) wrote about this being a marvelous upgrade to the 5532, but I cold be mistaken. Mouser and Digikey no longer carry it but I found some at ICPlus. Keep in mind that the 5535 does not have the current gain capability of the 5532, so if you have very low impedance headphones, this may not be a good choice, however, several Burr Brown

In any event, once you have socketed the op amp you should be able to find a long list or audiophile recommended parts that will fit, including some who make adapter boards so you can use some of the very expensive Burr Brown single op amps in the socket. Being able to hear the effect different op amps have may alone be a reason to mod the EAR. Here is a good resource for op amp replacements: Tangentsoft Op Amp List

Also, take a look at the op amp pins. I believe there are a couple of unused part locations which were meant to be filter caps but are now marked N/C. You can tell because one of the cap pins will connect to pin 4 or pin 8 of the op amp. If this is true, put a 10uF to 22uF Pana FM cap in each. Pin 4 should connect to the capacitors (-) side. Pin 8 to the (+) side.

Final Words
Like I've said, the Rega EAR is not worth purchasing for the sake of upgrading it. If on the other hand you already have it, you want to learn how to do things yourself, and you are happy to spend about $30 in parts in teaching yourself and your ears how different components affect your sound, then you are in an excellent place. Please let me know how your experiments turn out.

Thursday, November 23, 2006

Rega EAR Headphone Amplifier, Final Analysis

After taking a close look at the insides (see previous post) and listening to the Rega EAR headphone amplifier for about 3 hours I can finally give you my reader my definitive take on the product.

I tried four different pairs of headphones:

  • AKG K 701
  • AKG K 240 Studio
  • Shure E4C
  • Sony MDR-V600

and 2 different source components. I compared the EAR to my SUMO Athena non-headphone enabled preamp, as well as to the direct outputs of my Creative Labs Jukebox Zen 2.0. I actually chose the Athena because I knew that it had limited low impedance drive capability. I figured that it would emulate what a poorly designed headphone jack would do to the sound. On the plus side, it's pure class A and has about a 600 Ohm output impedance, which is pretty darn low for a line level amplifier. Still, it should have been blown away by the EAR.

After a few hours of listening, and after looking closely at the insides I can't recommend this product. Either in absolute terms or price / performance wise.

Compared to the Sumo Athena, it adds too little. The only times I could tell it was a better driver for the headphones was with the 29 Ohm E4C's. Compared to the direct outs, it seemed to muddle the sounds and rob bass drums of their midrange strike and impact. With the AKG 701s it seems to get harsh when there are complicated passages with a lot of upper midrange. Overall, it's a noisy piece of gear, which makes sense given the limited filtering done in the power supply. You can't turn it up too much before the hum of the power supply is clearly evident. The Sony MDR-V600s made the noise problem even worse. They are the most sensitive headphones I have and have a lot more treble than the AKGs which made the hiss from the Rega almost unberaable.

My sample even has some weird oscilation going on when you turned the volume up just a bit, which would disapear, but come on, guys. It's not like you don't have anyone in England who can show you how to design a current amplifier stage that won't oscillate. Otherwise I found the sound to be a complete "Why bother?"

My suggestion to would be amp buyers is to give this one a pass, or buy one, then sell the plastic Rega label on the front so you can go through it top to bottom and mod it until it shines. If. Well, never mind. Don't buy it.

Wednesday, November 22, 2006

The Rega EAR Gutted!

Last weekend after finishing a mid-term early, I decided to reward myself with a new headphone amplifier. I'd been using this as my primary listening station:

What you see are a pair of AKG K240 600 Ohm headphones. Why 600 Ohms? Because I'm driving them directly from the James Bonjiorno designed SUMO Athena preamp. It's a Class A design with tremendously low output impedance. Certainly low enough to drive these headphones, but it does get noisy if I turn up the gain as much as I need to.

I wanted a headphone amp so I could get gain and power, and free up some desk space. The Rega EAR is less than half the width of my preamp, and fits nicely in a cubby in the desk.

After getting the EAR home, the first thing I did, which was probably a mistake, was take it apart so I could look at the guts. In large part because no magazine review I'd seen had shown us the insides. And now I know why. Tis a sad thing indeed. I am hesitant to even show you but since I know thats why you come here, I'll give you the full naked view first:

From what I can see, you could make something better with about $40 worth of parts. $45 if you include the wall wart. Throw in another $40 and you could have top quality jacks.

The inadequacies with the price/value of this product begin in the power supply. It is single ended power supply, with a single voltage regulator. Remember this, because this is going to cause us problems in the audio section design, which we will cover later.

All of the electrolytic caps are standard grade Korean made Samwha capacitors. By the way, nothing wrong with Korean products. Especially their shoes! My complaint is that they are very standard grade parts, and that by using the simplest possible power supply design you end up having to route your output signal through capacitors before feeding your cans.

Another sad parts about this product is the single input buffer and possibly voltage gain stage. Rega chose to use a single op amp, the NE5532. It's a fine Op Amp. Phillips used to put them in their top of the line CD players, circa 1989. Come on fellows, it's 2006, almost 2007. Would it have killed you to at least use the class A biased NE5535 instead, or any of the superior Burr Brown or Analog Devices products out there today? And how about two of them so you minimize the crosstalk between them?

The one bit of good news is the discrete output stage. Unfortunately whatever good this does for us is undone by the power supply. Because it's a single ended supply, this means that there must be output capacitors. And there are. Two big electrolytics, which are NOT bypassed. That's right. No bypassing. Just straight through. Are they kidding me? No, seriously, an NE5532 and tin can electrolytic output caps? Are they bleeping kidding me?

The last problem I had was that the circuit board is single sided. There's nothing wrong with this, in general, but when you use a single sided circuit board you end up sacrificing the cleanness of the copper layout in order to save a few bucks in the board manufacture. This tends to affect the critical ground and power planes the most. Also, this is a dirt cheap thing to change.

If you can get this thing cheap, it may be a good started uprgrade project for some one. Rip out the electrolytics and replace them all with Panasonic FM. Bypass the output caps, and socket the Op Amp so you can experiment with a variety of op amps.

I did get a chance to audition the Rega EAR at the store with an AKG 701, and compare it to a pair of 600 Ohm AKG 240s. In my opinion, the AKG 701's started to sound compressed and tired when music got complicated and loud.

Over the rest of the week I'll start actually listening to this thing in my system. And maybe after that I'll start discussing the upgrades. But seriously, it's almost a sad thing to NOT upgrade this thing.

Based purely on the parts they put into this product, I have to say, you are better off saving your money and getting something like the HeadRoom Micro Amp for $299. It has seriously better parts value than this... throw back to the days of early mass market Japanese components.

Sorry Rega, I think that this is an exorbitant amount of money to charge people for what you are giving them. Get rid of the extruded aluminum exterior, and take the money you'd save on it and put it towards some decent grade components.

By this Thanksgiving weekend I'll have my actual listening tests concluded and will publish the results.

Sunday, November 19, 2006

Coming up Next: The Rega EAR

Gentle Reader,

I have purchased a Rega EAR headphone amplifier for use in my ever evolving home office system. It's gotten great reviews among the head can users plus I got it at a steep discount for being sitting too long as the demo unit at a local headphone store. Don't let that fool you though, most of the other one's disapeared quickly.

Over the next few days I'll be evaluating it against my reference standard headphone amp, a vintage Sumo Athena, as well as the direct outputs of my Creative Zen 2.0 driving my AKG 240 600 Ohm cans as well as the Shure E4C's.

I have to tell you, construction wise, I think it's very little product for the money. $299 new is a lot of change to put down on a circuit board with so few parts, that uses an off the shelf wall wart as it's primary supply. I'll be posting pictures of the insides, and recommended upgrades as I go along, if I think it's worth it.


Sunday, November 12, 2006

Man, Just how many Transistors Can you Replace?

It's been hard going here folks. The 25V supply is not showing me any love, but I am getting closer. The most likely problem I'm having is the darlington arrangement. The T0-92 which drives the TO-220 shorted, and at some point the brand new 6.8V zener I put in also smoked.

I have found upgrades for most of the NPN and PNP transistors used, usually just an improvement in the voltage rating. This may not sound like anything useful, but actually, since I don't have a load on the power supply, the voltage after the bridge rectifiers is unusually hot. Higher voltage transistors, and a higher wattage zener will help me keep the thing alive longer until I can put it back into the tuner.

After this, it's back to the 5V supply. Which is basically where I was a month ago. :))

Saturday, November 11, 2006

Beware the Lurkers!

One thing I've come to find out about you, dear readers, is that when my blog posts positive things, I get a lot of requests for more information, maybe even some custom mods. But when I post that I've just smoked a resistor, people who were previously sending me e-mails suddenly pretend they don't know me.

So from now on, it's nothing but good stories! About how wonderful the tuner sounds now that I've modded it, and how incredibly sexy the insides and outsides look, and how the mods I have made to my Conrad Johnson PV-10/Erik surpass all tube preamps, ever. All lies, of course, since the tuner isn't even working, and the PV-Erik sounds good, but not better than the good folks at CJ do on their better days, but come on, what do you want? Endless florid reviews that amazingly all say that every single piece of equipment I listen to is priced exactly what it is worth? Don't you get enough of that from the audiophile rags? Well, OK then, I can write like that too, and from now on, I will.

Psych!! Haha! This is not one of those sites I'm afraid. This blog is for the Solder Slapping Adventurer! The Geek who wields his iron like a light saber, striking fear into the hearts of tweak peddlers, gullible consumers, and sometimes our own children and spouses. We are not made of fearful stuff, for we are warriors and we undestand that the feeling of a 500V capacitor discharging through our fingers then down the left side of our rib cage to grab the pericardium so hard we can feel it's entire circumference around our hearts is to be knighted by Thor himself! Now raise your temperature controlled soldering iron into the heart of the storm above your heads, into the toxic fumes of evaporated rosin and scream your defiant battle cry:




Sunday, November 5, 2006

NPN's, aren't they all the same???

Well, kind of sort of. :) I think I found a temporary solution to my blown transistors issue, an MPS A42 which thank goodness does not have a center collector. It's also a 300V part, but what the heck, I'll take whatever I can. The gain is close to the originals.

This may all be moot, as I have wasted most of the weekend and haven't really studied. I really wanted to listen to my tuner before the end of the week. Guess it will have to wait.

Saturday, November 4, 2006

Honey, what's that smell?

"Um, nothing dear.........."

I replied, pulling the power cord on the power supply I was trying to revive.

I had the power supply 90% operational, the only thing I was having issues with was too little on the 5V supply. So, I was re-checking the the voltages using little pieces of wire in the Molex jacks, when I accidentally shorted the 25V supply go ground. Urgh!! !

That smell which wasn't really there was the bouquet of a young 5 watt Mills resister starting to melt it's protective cover.

Now, the PS is 30% operational. I have 15V, but no 25V or 5V coming out of this thing. I could probably get it working again with some generic replacements, but for some reason I don't have a single NPN transistor in a TO-92 case in the house. I guess it's time to order from Mouser again, and I might as well order more power resistors from Parts Exchange, if I'm going to be smoking what I have.

I can't find a source for the BD 419. I can find close matches, but most of the modern replacements have a center collector, which make them hard to retrofit. At this point I feel lucky just to have found the general specs for it. In case anyone can help find a replacment I am looking for:

  • Pins: EBC
  • Power: 2W
  • Hfe: 60
  • Vce 100
  • Through hole, so TO-220 or TO-202

On the other hand this may be a sign from Eutectic, the God of Solder, to design a new power supply board. I may very well end up with the single most expensive Tandberg 3001 ever. :)

More later!