What about your rule "Don't turn it on - take it apart!"?? I'd love to see the PCB, Transformer, Input protection and so on.
What about your rule "Don't turn it on - take it apart!"?? I'd love to see the PCB, Transformer, input protection and so on. ;-)
Its way too long for single episode
Even its 1 or 1.5 hours is still fine, IMO.Its way too long for single episode
I thought this episode was also way too long, and there is already some stuff left out. I didn't want to spend the time to edit and refine further.
The teardown is an additional 45min, and includes some extra testing. I had to re-shoot almost the entire thing, wait for the rant in the video...
Dave.
so to kill the gap, was it high side or low side current measurement, or some combination?
Teardown is already uploaded, coming tomorrow. 45min worth!It's a good thing you 'straighans live in the future, so Tuesday typically means "late Monday" to the rest of the world.
Am I the only one who's really, really horrified by the display? As in, "no way in hell would I ever buy it" horrified? The almost-white-on-slightly-less-white is ghastly. I think I would spend as much time squinting at it as I do operating it!
Am I the only one who's really, really horrified by the display? As in, "no way in hell would I ever buy it" horrified?
... wait for the rant in the video...
first mode would be ripping that bloody beeper.
Apart from the UI seems like a great product for the price.
Dave,He needs 30000 count at that moment, does the 87V have it? Nope. And this is 2013, the 87V is very very stuck in its times :-//
I see you're favoring the Agilent DMM over the '87V more often...
A new love?
Too bad. I was looking for another power supply but with the shitty controls it is worthless. As others have said the user interface can cause some real problems. Sorry Dave but I am disappointed that you accept this as not a fail.
I suspect this PS will be one of those a lab will buy and it'll sit in the corner of the lab ... 'til it's sold at a surplus sale.
"Have you ever used the Atten?"
"I tried once but couldn't figure it out."
He needs 30000 count at that moment, does the 87V have it? Nope. And this is 2013, the 87V is very very stuck in its times :-//
Dave, did you had a peek inside before the review?
What did Dave say in the intro "a pretty high-end bit of kit"? That train wreck passes as high-end? Muhahahahaha :-DD :-DD :-DD
I can think of quite a good project for anyone with some spare time on their hands, though. The hardware in the box seems OK, and the whole thing is serial programmable, so I wonder whether it might be possible to build a dedicated external controller for this thing with a better LCD and front panel controls? It could be quite a good "my first PIC" project; rather than building a power supply from scratch, just make a relatively inexpensive off-the-shelf one usable.
Good review. Personally I'd actually give that a fail. Because even if the hardware is ok, and it has those good features, it's so quirky and difficult to use that it's almost guaranteed to eventually cause you - or a colleague - to make a mistake which sends it into overvoltage or overcurrent and blow the arse out of your project. It's bordering on dangerous, except that it only outputs 30v.
Its core abilities at the price look quite much and it seems basically delivering them safely. When we approach an $500 outlay for a PSU we need it be a product though, don't we? Can a non supporting brand which not even replies a simple email query and has the basic PC remote control thing non functional as it comes be considered as offering a product at the price or just a "shipment"? If it was to be well functioning with nice PC software, having good support, warranty to be expected honored etc. and would cost how much? $750-1000, which ones are the offerings it should really be compared to? Something from TTI maybe?
With a power supply, however, the designers are the same people as the target market! What is going on?
I had a look inside. The separate serial board goes to an MCU board, centered around a chip labelled STC 89C58RD+. A quick google turned up a datasheet that contains a few package outlines and no text whatsoever. Good luck getting it to work by reverse engineering that.STC 89C58, standard 8051-based MCU. STC is a Chinese manufacturer.
The OVP feature can actually make it a safer supply than better higher prices supplies that don't have that feature.Do you know how they have implemented the OVP? The traditional OVP circuits are independent from the rest of the power supply circuitry so they can protect the load from power supply failure. In the case of shorted regulator transistors, the OVP is meant to protect the load by forcing the power supply's fuse to blow. If the OVP level is set by a DAC controlled directly by the main processor, it is not independent.
Do they have a separate controller and separate supply just for the OVP?
Too bad. I was looking for another power supply but with the shitty controls it is worthless. As others have said the user interface can cause some real problems. Sorry Dave but I am disappointed that you accept this as not a fail.
When did I say it's not a fail? I spent maybe 30min complaining about the bloody thing!
Will I personally continue to use it, well, yes, I've persevere because it's the most versatile supply in my lab. Just like I find my Fluke 87V defaulting to AC current incredibly annoying, but I continue to use it.
Too bad. I was looking for another power supply but with the shitty controls it is worthless. As others have said the user interface can cause some real problems. Sorry Dave but I am disappointed that you accept this as not a fail.
When did I say it's not a fail? I spent maybe 30min complaining about the bloody thing!
Will I personally continue to use it, well, yes, I've persevere because it's the most versatile supply in my lab. Just like I find my Fluke 87V defaulting to AC current incredibly annoying, but I continue to use it.
Yes you complained and then gave it a thumb sideways. I took that as not a fail overall. I do see your point about the actual output being good but its interface lets it down to the point of total failure, IMHO only.
This (http://bbs.dianyuan.com/bbs/u/82/1047621254556883.pdf) appears to be the Chinese datasheet for it.Thanks! The URL on that PDF led me to the English version (http://www.mcu-memory.com/datasheet/stc/STC-AD-PDF/STC89C58RD+-english.pdf). It has ISP functionality, but it's not immediately obvious to me if there's a way to dump the existing firmware in the MCU's flash out to the computer though. I'll need that if I'm ever going to get a list of serial commands (as Dave was not forthcoming).
Their programming software doesn't appear to have a "verify" function either, so I guess ISP won't do it. But it's a standard-looking package 89c58... perhaps holding it in reset and doing the usual thing on an 8051 parallel programmer would work (if security bits aren't set). At least see if you can read its ID that way. Maybe STC cloned some other manufacturer's 89c5x design and just wrote an ISP bootloader to it.This (http://bbs.dianyuan.com/bbs/u/82/1047621254556883.pdf) appears to be the Chinese datasheet for it.Thanks! The URL on that PDF led me to the English version (http://www.mcu-memory.com/datasheet/stc/STC-AD-PDF/STC89C58RD+-english.pdf). It has ISP functionality, but it's not immediately obvious to me if there's a way to dump the existing firmware in the MCU's flash out to the computer though. I'll need that if I'm ever going to get a list of serial commands (as Dave was not forthcoming).
This can only work when you protect your design by custmized hardware, when nobody else can get the specificBut cloning the firmware without open source is impossible? What a strange argument, if they want to clone you they will ... period.
chip/asic they you can publish any kind of info and work with open source ppl (but then still someone, if there
is enough profit, will clone your custom chip).
Anyone that sent a email to atten trying to get details about the unit hear anything back yet? The software sucks and I would like to create something a lot nicer...No, they didn't reply. However they replied to TME, where I bought it, asking which browser I was using.
function interface_parts = Atten_controller_interface()
%% connect to the device
Atten_conn = serial('COM1','DataBits',8);
interface_parts.Atten_conn = Atten_conn;
fopen(Atten_conn);
interface_parts.is_connected = true;
interface_parts.do_shutdown = false;
%% create GUI
%GUI window
interface_parts.figure = figure('Position',[100,100,800,600], 'CloseRequestFcn', @Atten_disconnect);
%Create plot for values
interface_parts.data_plot = axes('Units','pixels', 'Position', [100 150 600 400]);
%buttons to toggle the device on/off and to exit
interface_parts.output_btn = uicontrol('Style','togglebutton', 'String','output','Position',[350,10,100,50]);
interface_parts.exit_btn = uicontrol('Style','togglebutton', 'String','exit','Position',[475,10,100,50], 'Callback', @Atten_disconnect);
%labels and text boxes for the user to input desired V and i values
interface_parts.V_label = uicontrol('Style','text', 'String','V','Position',[150,10,50,20]);
interface_parts.V_field = uicontrol('Style','edit', 'String','0','Position',[200,10,50,20]);
interface_parts.i_label = uicontrol('Style','text', 'String','i','Position',[150,40,50,20]);
interface_parts.i_field = uicontrol('Style','edit', 'String','0','Position',[200,40,50,20]);
%% Initialise the hex vector
interface_parts.control_vector = ['AA'; '20'; '00'; '00'; '00'; '00'; '00'; '00'; '00'; '00'; '00'; '00'; '00'; '00'; '01'; '00'; '01'; '00'; '00'; '00'; '00'; '00'; '00'; '68'];
interface_parts.default_vec = interface_parts.control_vector;
vals = []; % Initialise the array for measured values
tic; % Start timer
%% keeping going until the user exits
while(true && interface_parts.is_connected)
Atten_apply_vector(1,2); % set device
[input_dec,count,msg] = fread(Atten_conn,24,'uint8'); %read status
%convert and store results
V_val = hex2dec([dec2hex(input_dec(3)) dec2hex(input_dec(4))]) / 100;
i_val = hex2dec([dec2hex(input_dec(5)) dec2hex(input_dec(6))]) / 1000;
vals = [vals; toc V_val i_val];
%show the results
[AX,H1,H2] = plotyy(vals(:,1), vals(:,2), vals(:,1), vals(:,3));
xlabel('time [s]');
set(get(AX(1),'Ylabel'),'String','Potential [V]');
set(get(AX(2),'Ylabel'),'String','Current [A]');
pause(0.05);
end
function Atten_apply_vector(hObject, eventdata)
%note that matlab uses an indexing system starting from 1, whereas
%the protocol pdf assumes an indexing system starting from 0
%process the voltage value chosen by user
V_val = str2num(get(interface_parts.V_field, 'String'));
V_dec = round(V_val * 100);
V_hex = dec2hex(V_dec, 4);
interface_parts.control_vector(3,:) = V_hex(1:2); interface_parts.control_vector(4,:) = V_hex(3:4);
%process the current value chosen by user
i_val = str2num(get(interface_parts.i_field, 'String'));
i_dec = round(i_val * 1000);
i_hex = dec2hex(i_dec, 4);
interface_parts.control_vector(5,:) = i_hex(1:2); interface_parts.control_vector(6,:) = i_hex(3:4);
%update output state
interface_parts.control_vector(16,:) = dec2hex(get(interface_parts.output_btn, 'Value'));
%send suitable vector to the device
if(interface_parts.do_shutdown) %set the device back to the default state before disconnecting
interface_parts.store_control_vector = interface_parts.control_vector;
interface_parts.control_vector = interface_parts.default_vec;
end
output_dec = hex2dec(interface_parts.control_vector);
fwrite(Atten_conn,output_dec,'uint8');
end
function Atten_disconnect(a,b)
if(interface_parts.is_connected)
%shutdown system
interface_parts.do_shutdown = true;
Atten_apply_vector(a,b);
%close connection
fclose(Atten_conn);
delete(Atten_conn);
clear Atten_conn;
display('disconnecting from device');
interface_parts.is_connected = false;
else
display('disconnecting... not connected to device');
end
delete(interface_parts.figure);
end
end
It doesn't. The serial port is hardcoded to COM1.
Welcome to Atten. Leave your brain at the door; we'd like you to write some software for us.
In any case you can fix the assigned COM port number to a specific FTDI chip (by serial number), see here (http://www.ftdichip.com/Documents/AppNotes/AN_132_Re-Assigning_COM_Port_Numbers_Using_Registry.pdf). That got the software working, and I'm duly seeing lots of protocol data. I'll have to see about getting that documented.
Sweet! got a replay with sample code (matlab maybe) and specs on the communication protocol.Excellent! I notice the document is marked atten.eu -- that's the European reseller. I've bought an Atten ATZ97 from them, and that came with a great manual, including complete protocol documentation. I'm pretty sure they wrote these themselves. Shouldn't have cheaped out and bought from TME, clearly.
Sorry for bringing this old post back up, but I was wondering if anyone changed out the fan because the stock fan is noisy as hell.