• These commercial threads are for private transactions. diyAudio.com provides these forums for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members, use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

Lovoltech LU1014D power JFET

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Some people frequent the Pass Labs Forum and some don't.
That's okay--it takes all kinds to make the world go 'round. We don't hate you...promise.
--but--
I will now give you an excuse to visit Nelson Pass's DIY site (www.passdiy.com). You see, there's a device that few, if any, of you have in your arsenal. In fact, very few people even know that such a thing exists. It's called a power JFET. Sure, you're aware of JFETs. Nifty little things. High Zin--close enough to infinite for all practical purposes, and quiet if you're into voltage noise (separate and distinct from current noise--bipolar fans go grumble somewhere else). Wonderful devices. Marvellous. Until you find that it's difficult to get more than a few milliamps through them.
Bummer.
The audio gods have heard your cries for help and have responded with the LU1014D.
The first thing you want to do is go to www.passdiy.com and read the Zen Variations 8 paper that Nelson wrote. Then you go to the manufacturer's website (www.lovoltech.com) and look at the datasheet for the LU1014D. It wouldn't hurt to look at a couple of the threads in the Pass Labs Forum for other hints about using these devices. Nelson already has a commercial design out using the LU1014D and more coming. If any of you are brave enough to build something I've cobbled together, I'll be posting a higher powered amplifier soon. It is my understanding that Nelson also has more diy-oriented projects in the works.
If you find that you want to build the Zen #8 or perhaps some other design, your next instinctive move would be to hit Mouser or Digikey. To no avail. Unfortunately, Lovoltech does not have a retail distribution network at this time. They're geared for the computer motherboard makers and sell directly to them. Little guys like us are left facing a rather hefty buy-in requirement (well over 2000 devices, minimum order). Uh, oh.
A group buy was initiated, but never quite developed critical mass. I, not wanting to wait for all the details to get sorted out, employed the Gordian Knot solution. I bought a bunch of the silly things out of pocket so I could get started on my own designs. In fact, I bought quite a lot of them.
I am offering the Lovoltech LU1014D power JFET at the following prices:

Unmatched
1 $1.69
20 $1.49
100 $1.39
200 $1.33
1000 $1.29

Matched (Vgs)
1-19 $2.59
20-50 $2.29
(Anybody who wants more than 50 matched devices needs their head examined.)

This is how it works--
--All prices in US dollars (not Australian, Canadian or other dollars). Personal checks okay. When your check clears the bank, the JFETs ship. My bank is the final arbiter as to whether a check is good or not. Sorry, no PayPal or credit cards.
--No minimum order requirement. You want 1, you get 1. You want 37, you get 37. You get the quantity you order. It's that simple.
--No handling fees or other oddities. The price listed above is it.
--Shipping will be pretty much at cost and as cheap as we can arrange it. For US (and possibly Canadian & Mexican) buyers, I can mail parts in a standard business envelope with minimal padding. They're small, fairly robust, and lightweight. This means that an average buyer will be looking at a few cents for an index card to tape the things to, a few cents for an envelope, and $0.39 for US postage. If you want the added cushioning of a padded envelope, that will probably add something like $0.75 to shipping. All told, an ordinary buyer here in the US will be looking at about $0.50 shipping if you go the envelope route. Nearly everyone will get out for less than a buck.
As I've said elsewhere, these things are about the size of your pinky fingernail. Even people who want 100 devices should fit into the envelope category.
I will try to accomodate anyone who wants UPS, FedEx, carrier pigeon, etc. but please try to be reasonable. The weirder it gets, the more hassle it will be for me, and the more it will cost you.
--No refunds. If the parts get smashed in shipment, return them and I will send replacement parts. If you blow them up and try to return them, I will hire a Voodoo priest to make your cat miserable.
--I will handle North America. David Briggs (dhole) has offered to handle European distribution. If you live in Europe, talk to him. No other arrangements have been made yet.
If, brave soul, you have made it through all above and haven't lost heart, this is what you do:
Tag the e-mail button at the bottom of this post and give me
1) Your DIY userid.
2) Your real name.
3) How many you want.
4) Whether you want them matched or unmatched.
I will respond with an e-mail giving you my address and a total, including shipping. Please do not pass on my contact information to anyone. If you've got a buddy who wants parts, get them to sign up here or order for both of you in your name.
Please be aware that matching takes time. I will almost certainly be slower to ship matched parts than unmatched. Be patient if you want matched devices. I promise to get them out as soon as possible.
Note that a post in this thread does not qualify as an order. Only a direct e-mail to me qualifies as an order.
If you have questions...ask. Either here or in the Lovoltech group buy thread in the Pass Labs Forum.

Your (more or less) Humble Servant,
Grey
 
This is a blatant cut-'n-paste from the other thread. I'm trying to save time for answering e-mails.
Oh dearie, dearie, me...
The floodgates have opened and lemme tell ya somethin'--the Mongol Hordes could learn a thing or two from DIY folks hungry for JFETs.
Okay, a little housekeeping. I'm getting e-mails that aren't including the basic info I requested:
1) Your DIY userid
2) Your real name
3) Your address
4) How many devices you want
5) Matched or unmatched
I even had one that was just a name, but none of the other info. No quantity, no address, nothing.
Sigh.
I've put the baby down for a nap (that's a joke, son, two year-olds don't sleep...ever) and am going to try to get a few e-mail replies out. I will attempt to get to everyone who has ordered so far, but don't panic if it's a little later in the day, or perhaps even tomorrow.
I'm fine on stock. You want 'em? Send me an order. I don't anticipate running out until 5 this afternoon. (That was another joke. Don't panic. I bought plenty.)
I've had an order from Indonesia and one from Australia. Technically, I'm not handling orders from that corner of the globe. I'm doing North America (Mexico, US, Canada) and David Briggs (dhole) is doing Europe. No one else has stepped forward to handle other areas yet. Now, if it doesn't get any worse than one order each, I may go ahead and deal with it, but I've always had the impression that Australia had more DIYers than the rest of the world combined. With that in mind, I anticipate more Aussie orders.
Let's do this--if you live somewhere other than North America or Europe, go ahead and send me an order. Warning: For the time being, I'm just going to let them pile up. Do NOT anticipate immediate shipment. If someone steps forward to handle a given area, then I'll forward the orders to them. If not, maybe I'll open a bottle of Lagavulin and try to work it through myself. We'll see. One way or another, we'll try to get JFETs to everyone who wants them, even if it takes a while. Sorry, but there's only one of me and I'm running on two hours sleep, which is scant even by my standards.
I've had one fella ask if it was okay if he ordered through a US member (he lives elsewhere in the world). Absolutely! I'm just trying to stay out of all the paperwork and junk involved in shipping overseas. If you live in, say, Peru (say hello to Paddington Bear for me), and have a buddy here in the US, then feel free to make arrangements with them. I'll ship to them, and they can ship to you. If you're happy with that, then it's fine by me.
Quick! Somebody fill my glass!

Grey
 
Another cut-'n-paste...
If--and only if--you live in the US, you should have return e-mail from me by now. Given that I'm running on two hour's sleep and lotsa caffeine, double-check my math. If I did it wrong, please accept my apologies and drop me a line. We'll fix it.
Steen, you're included in the US batch. If you and your surrogate have not gotten e-mail, let me know.
I'll be tackling Canada later tonight or (more likely) tomorrow, but first I've got to research postal rates betwixt here and there.
Those from other corners of the globe, sit tight.

Grey
 
Another update:
Folks, as a matter of both statistics and logistics, I need you to think in terms of smaller groupings for matched devices; perhaps 8, 10, or 12. I have people requesting very, very large single groups and it's approaching the point where I will have to strip the entire population of parts of certain values in order to make up the matched lots. Obviously, that's not fair to the members who want to match their own--they'll end up with holes in their values.
My intention is to try to hit 10mV windows, i.e. +-5mV. There may be one or two millivolts drift to either side, but that's my goal. If you approach it from a background in statistics--or if you refer to Nelson's Vgs curve posted elsewhere and work it through that way--you'll soon see that if someone asks for, say, 100 matched parts, it's going to require going through a huge number of individual units to meet that request. Huge meaning thousands in some cases.
Incidentally, I'm getting a slightly different distribution from the one Nelson posted. On his graph, there's a shorter bar to the left of the peak of the bell curve. Mine is just to the right, just doing an eyeball survey of the quantities I've graded so far.
Also note that these parts are not behaving the way tubes of IRF MOSFETs do. There's not a more-or-less smooth gradient going from one end of the tube to the other. I'll have a high value, then a low one, then perhaps as many as three or four that cluster towards the center, then back to another low one. Scattershot. I had one tube that had a notable preponderance of lower values, followed later on by another tube that leaned heavily towards higher ones. One tube clustered more towards the middle. The rest have been all over the map.
The devices do not have stamper numbers that might hint where they originated in the original wafer. I do not know what their manufacturing process is, nor do I know how they move their parts from manufacturing to packing. Presumably it involves a fair amount of jumbling around, judging from the value trends I'm seeing. Or lack of same.
Anyway, please think in terms of smaller sets. It'll make life easier for your fellow DIYer.
(Not to mention me...)

Grey
 
And an update to the update:
My previous post is not meant to discourage you from ordering matched devices. Just plan on something like two sets of 10 or four sets of 12.
Yes, I know it's difficult to guess what you'll need when Nelson's Zen #8 only needs 1 per channel and I haven't posted my somewhat larger design yet. As guidance, think of an Aleph, where the active devices on top need to be matched as one group and the ones on the bottom need to be matched as a group...but they do not need to be matched together.

Grey
 
GRollins said:
Mark Stones (mark stones, here on DIY) has offered to take orders for Australia and 'and the odd one or two orders overseas.'
I'm on a short leash at the moment--got to move on.
Thanks, Mark.

Grey

As Grey has pointed out I will be doing the Australia orders and one or two overseas orders that are around this side of the planet, unlike grey I will be doing this only once so if you are interested Email me your requirement prior to the 5th of February.

The prices will be whatever grey charges + postage + envelopes + tax (if I get charged then so do you) + share of bank transfer fee (for money to pay grey) + paypal if you pay me with that medium.

In short you will pay what it costs me nothing more hopefully nothing less.:xeye:
 
IMPORTANT: PLEASE READ IF YOU ARE ORDERING FROM CANADA
Without warning, my bank is charging a $16.75 fee for processing checks from Canada. This is a flat fee, regardless of the amount of the check. They say that it is a fee imposed on them by the Fed and that there's nothing they can do about it.
I have expressed great displeasure at having this sprung on us with no warning.
I don't yet know about Mexico, but checks from other countries will be more. It will probably be cheaper to do a wire transfer.
If you have not yet sent me a check, then add $16.75 to the total. Since this is a flat rate addition to the amount, you might consider stocking up and writing one larger check, but that's not my decision to make; it's yours.

Grey
 
Update for Canadian buyers:
paulb has come up with a suggestion that seems as though it will work. Get an international postal money order drawn in US$. We have been told that as long as it is in US$ it will not be subject to this absurd fee.
I will not rest easy until I see this in action, but let's try it.
Until further notice, Canadian buyers should use international postal money orders, not personal checks.
Those who have sent personal checks, don't panic. You did so in good faith. I will attempt to find other ways to avoid this fee.

Grey
 
Update to the update to the Canadian update...(or something like that)
I am now told that the international postal money order needs to have an "ABA" number on it. (Stands for: American Banking Association[?]) Appearently many checks already have it, but don't assume.
I got some matching done this afternoon and am headed down to the dungeon to take up where I left off.

Grey
 
canadian orders

Hi, Grey;

You may want to check your closest full-service post office for processing fees with money orders (that's what my order is). I have been using this specific method for paying US orders for the last 20 years and had very little problems with it. Our postal systems have had an information mechanism in place for longer than I can remember for doing these kind of transactions.

BTW, I sent my order last Monday so let me know, at your convienence, how it works out trying to cash it. I included a little more in case of a small extra expense on your end. Thanks for all your effort in this, it is greatly appreciated.

Todd.
 
This is where things stand.
I have rough-graded a little over a thousand devices. I have gone back through and refined the groupings.
For those of you who have been worrying about whether I was going to have to rape the entire population of devices in order to meet matching requests, fret no more.
With that starting population of devices, I was able to meet ALL outstanding requests for large groups. In fact, I can still meet a few more requests for large groups. Just a few--so don't overdo it, okay?
However...
I am going to reiterate what I said in the other thread (and possibly in this one--I kinda got lost as to where I was in the posting sometimes): Please be reasonable in requesting groups of matched devices. I would like to ask that you consider 24 devices or less...preferably 12 or less.
Note #1) This does not mean that I will shoot you if you ask for 48 devices in one group. It means that I reserve the right to tell you no. For that matter, the time may come when I have to tell people requesting groups of 24 that I can only give them 16, or whatever. Please try to take the news gracefully.
Note #2) I will be happy to meet requests for larger numbers of matched devices...as long as you are willing to accept smaller groups within the total amount. You want 40 devices? No problem...as long as you are willing to take that quantity as 2x20 or 4x10, something like that. You want 100? Let's talk about 10x10...you get the idea. I promise to try to meet requests for single large groups, but I make no guarantees.
I repeat: I will not run through untouched tubes of devices just to cobble together a single group of 100 matched devices for someone. It's not fair to other DIYers.
Be reasonable. As someone noted in an e-mail--you can't use that many in a single circuit anyway.
I still have unopened bags with unopened tubes. If you want 'em, I've got 'em. I am on the verge of opening another bag, so don't be shy.
At some point, I reserve the right to go ahead and match all the remaining devices. I may. I may not. If I'm going to do so, I will issue a "last call" for unmatched devices. After that, only matched devices will be available. Trust me, I'm in no mood to do so today, tomorrow, or even next month. This is tedious work, and I'd like to get back to the circuit I've been working on.
Since there's only one of me, and I've been working on getting devices matched, I have not been answering much of my e-mail. I have put a priority on getting people in touch with Mark Stones (who is handling Australia, Indonesia, and other such places), since he has a cutoff date. If you wrote me sometime in the last 36 hours or so, please be patient. I will get to you.
People need to be a little more attentive to the information I'm requesting. For instance, I can't mail your devices to your e-mail address. I need a physical address in the real world.
Again, to order press the little e-mail button at the bottom of this post. I will need the following information:
1) Your DIY name
2) Your real name
3) Your address
4) How many devices you want
5) Matched or unmatched?
I had a whole laundry list of things in mind that I wanted to hit in this post. I think that's most, if not all of it.
By the way, thanks, guys. I'm gratified by the response.

Grey

P.S.: One poor soul thought I hated his guts. I had posted something in another thread that he thought referred to him in a derogatory manner. It didn't. If you think I'm mad at you, I'll make it easy on you...I'm not. It's been a long time since I've had a genuine problem with another member. There are people here that I ignore and others that I take with a grain of salt, but I'm not mad at anyone.
There, I think that covers it. If not, I'll get back in and post again.
 
Grey-
but I'm not mad at anyone.
------------------------------------------
Grey be good, why are you always upset with what others think or say, chill out man!!, we think you do a great job here, just
dont be so stressful to yourself

"If you think I'm mad at you,"
grey come on you cant be serious!, what on earth gave you
that idea chill out man thats all im going to say for now

you are the nicest person ;)
 
A little more information on what works and what doesn't regarding checks from Canada:
An "International Money Order" made out in US$ worked.
A Canadian check in US$ drawn on the Wells Fargo Bank here in the US worked. There wasn't anything on the check that identified it as being one particular kind of check vs. another, but it had an ABA number that made my bank happy.
A Canadian Postal Money Order didn't work...even though it was made out in US$. It lacked the magic ABA routing code and would have had to to be presented to the Fed for collection from a Canadian bank.
I still say this is absurd, but that's just me being...well...me.
Just for reference, this ABA thing is nine digits long. Some routing codes are 8 digits, some, it seems, are 4. There are others. It were too many for me, as the expression goes. Unfortunately, there doesn't seem to be a standardized place for it down in that string of numbers along the bottom of the check. On one check it was first. On another, it was third. I confess that I'm not clear as to how to tell which group of numbers it is. You would think that there would be standards, but apparently not. I'm sure that somewhere amongst our members we have someone who works at a bank who can explain this in plain English so that a po' dumb Southern boy can understand, but it seemed to me that everything they said was contradicted by something else they had said.
I am not sending checks through that will be subject to this ridiculous fee. I'll contact the sender and see how they want to proceed. Even if it costs a dollar or two to get another check cut, it will still be less than $16.75.
Suffice it to say that I'm mad all over again.
Sorry guys, I anticipated some difficulties, but not this one.

Grey
 
A quick update on the Canadian Postal Money Order. As Todd suggested above, the US Postal Service did honor it with no service charge.
Given that the USPS is a division of the US government, as is the Federal Reserve, this leaves begging the question of why the Fed charges a fee, but the USPS doesn't.
Or we could assume that my bank is fibbing a bit, although I've not had trouble with them in the twenty years or so that I've been with them.
Flip a coin. Make your own guesses.

Grey
 
I've got a few seconds...thought I'd drop in a few words about devices.
As I've noted before in one or another of these threads, I haven't done a formal analysis of the curve, but I'm getting the sense that these devices are all slightly to the left of Nelson's values. It belatedly occurred to me that it may very well be because I'm using a heatsink. I haven't had time to try to figure out anything like a 0 tempco for these things, but 1A ain't it. It's a positive tempco at that current. The numbers keep climbing as the device heats up.
Regarding distribution of values--I've seen nothing to change my earlier opinion that these things are all over the map. One tube tends towards high values, another towards low values, and the next acts like someone stirred all the values together in a pot. There's no rhyme nor reason to it. Those who are ordering unmatched devices take heed; you'll need to think in terms of more raw devices to find the same sort of matches that you might expect from IRF, OnSemi, et. al.
As far as absolute values go, I've seen numbers as low as .75V or so. On the upper end, it tapers off pretty quickly above about 1.14. Again, not having plotted a curve, I have a gut feeling that the curve is asymmetrical. The peak is slightly to the right of where you would expect the peak of a normal bell curve to be. The statistical term for this is skew if you feel motivated to look up such things.
(If you want an audio analogy, think of a bandpass filter at 1kHz with a 6dB/oct HP slope on the left and a 12dB/oct LP slope on the right.)
To those who end up with values out on the tail ends of the curve--don't worry about it. They'll work just fine for headphone amps, preamps...for that matter, even amps. If Nelson is going where I think he's going with the adjusting Rs to match the Vgs trick, you'll be using smaller resistances to bias low Vgs values. (Don't start ragging me about the relative increase of the device...that's another topic for another day.) If you're the sort who likes to reduce Rs in order to reduce the Zout of the amp, you'll be actively seeking low Vgs values. You guys can set up an entire secondary market swapping values back and forth.
I can see the ad now: "In search of eight .83 devices to match the ones I've got on hand. Willing to offer .91 and/or .96 devices in trade. Call 555-5555 after 8PM."

Grey
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.