you know there is a demo board for it direct from Ti right that you can buy and its already soldered? do you have tools for soldering this? you do realize its DFN package with a powerpad underneath so you NEED reflow oven or hotair tool to solder?
also I do wish people would not take advantage... he doesnt get this stuff for free, hes just really generous. fine if he puts feelers out for testers, but flat out asking... I dunno maybe you are asking to buy? in which case there are plenty of people ahead of you in the GB thread asking for when its ready
also I do wish people would not take advantage... he doesnt get this stuff for free, hes just really generous. fine if he puts feelers out for testers, but flat out asking... I dunno maybe you are asking to buy? in which case there are plenty of people ahead of you in the GB thread asking for when its ready
Last edited:
+1
Strongly agree. I hope many on this thread realize how advanced Ian's circuits are, and how fortunate we are to have a chance to play with this technology. The quality and speed he is designing is staggering, and on the top of that, maintaining pretty sizable group buys is amazing dedication. I think from one guy that is quite enough.
Thanks guys for your kind words. The pleasure you brought to me is more then what I did to you. Your feedbacks made my project meaningful.
Ian
AR2, qusp and Ian, you are all great members in diyaudio and always give great advise and support to members. Without you guys, I don't think I can finish this complicated project.
Ian, thanks for your generous to offer me the PCB.
qusp : I considered to buy the Ti evaluation kit, it's much cheaper than buying components but the size is a little bigger than expected. So I bought the adaptor. Thanks for reminding me that the cap have to be close to the pins, I will finish the prototype with SMD cap and see if it will smoke
Ian, thanks for your generous to offer me the PCB.
qusp : I considered to buy the Ti evaluation kit, it's much cheaper than buying components but the size is a little bigger than expected. So I bought the adaptor. Thanks for reminding me that the cap have to be close to the pins, I will finish the prototype with SMD cap and see if it will smoke
Last edited:
If I were you I would have a read of the datasheet for power dissipation and look at the load currents that you're expecting. Consider that you're also driving LCD etc from arduino.
You should be able to calculate if you expect smoke or not
qusp, I think with Ian's pcb you might just be able to get to the power pad from underneath. Though it's surely not going to be an easy (I mean heating from underneath while keeping the chip straight?) or even a neat or reliable job without a rework station.
What are you using to supply the regulator anyway? Seems an odd choice for the non-critical parts of the circuit. Considering that the i2c will be isolated and the relays are ... well it just feels odd to drive the coil of a relay with something with performance like this.
You should be able to calculate if you expect smoke or not
qusp, I think with Ian's pcb you might just be able to get to the power pad from underneath. Though it's surely not going to be an easy (I mean heating from underneath while keeping the chip straight?) or even a neat or reliable job without a rework station.
What are you using to supply the regulator anyway? Seems an odd choice for the non-critical parts of the circuit. Considering that the i2c will be isolated and the relays are ... well it just feels odd to drive the coil of a relay with something with performance like this.
Last edited:
If I were you I would have a read of the datasheet for power dissipation and look at the load currents that you're expecting. Consider that you're also driving LCD etc from arduino.
You should be able to calculate if you expect smoke or not
qusp, I think with Ian's pcb you might just be able to get to the power pad from underneath. Though it's surely not going to be an easy (I mean heating from underneath while keeping the chip straight?) or even a neat or reliable job without a rework station.
I've read the datasheet before and it recommend to use the copper on PCB as heat sink. I know the adaptor doesn't have enough copper for heatsink, I consider to attach a small heatsink on the chip to see if it help.
Pd = (Vout-Vin) x Iout (datasheet formula 6)
Tjunction = Tambient * [(theta J-A) * Pd] (datasheet formula 7)
With minimal pcb surface area (theta J-A) is roughly 120degC/W (off the end of the chart, if the thermal pad underneath that adapter has no copper at all to solder to then it might be more like 150degC/W.
Assume voltage drop across the LDO is 3V and max load current is 40mA and ambient inside enclosure of 35degC (with a few salas shunts nearby it could easily be more than this)
Pd = 3*0.04 = 0.12W
Tjunction = 35+120*0.12 = 50degC
That's probably realistic for driving one mechanical relay. Arduino and LCD might be more like 300mA (at a guess here).
In that case ... Tjunction = 143degC
Datasheet recommends 125degC as max junction temp for highest expected ambient and highest expected load.
See now how qusp and Ian think you'll have another visit from that dreaded blue smoke monster that lives inside these chips? There is no need for his visit to be a surprise, all the info is in the datasheet
Tjunction = Tambient * [(theta J-A) * Pd] (datasheet formula 7)
With minimal pcb surface area (theta J-A) is roughly 120degC/W (off the end of the chart, if the thermal pad underneath that adapter has no copper at all to solder to then it might be more like 150degC/W.
Assume voltage drop across the LDO is 3V and max load current is 40mA and ambient inside enclosure of 35degC (with a few salas shunts nearby it could easily be more than this)
Pd = 3*0.04 = 0.12W
Tjunction = 35+120*0.12 = 50degC
That's probably realistic for driving one mechanical relay. Arduino and LCD might be more like 300mA (at a guess here).
In that case ... Tjunction = 143degC
Datasheet recommends 125degC as max junction temp for highest expected ambient and highest expected load.
See now how qusp and Ian think you'll have another visit from that dreaded blue smoke monster that lives inside these chips? There is no need for his visit to be a surprise, all the info is in the datasheet
Last edited:
If I were you I would have a read of the datasheet for power dissipation and look at the load currents that you're expecting. Consider that you're also driving LCD etc from arduino.
You should be able to calculate if you expect smoke or not
qusp, I think with Ian's pcb you might just be able to get to the power pad from underneath. Though it's surely not going to be an easy (I mean heating from underneath while keeping the chip straight?) or even a neat or reliable job without a rework station.
What are you using to supply the regulator anyway? Seems an odd choice for the non-critical parts of the circuit. Considering that the i2c will be isolated and the relays are ... well it just feels odd to drive the coil of a relay with something with performance like this.
hochopeper,
My TPS7A4700 PCB has a hole under the QFN20 thermal pad
. So, it would be easier to solder the pad from back of the PCB even by a normal soldreing iron.Although it rated at 1A, I don't think it can keep good low noise performance for a output greater than 200ma. So, I don't suggest for a high current application
. Ian
Last edited:
hochopeper,
My TPS7A4700 PCB has a hole under the QFN20 thermal pad
Although it rated at 1A, I don't think it can keep good low noise performance for a output greater than 200ma. So, I don't suggest for a high current application
Ian
I thought I had seen that hole when looking at the prototype pcbs when you posted photos. I'm still not sure about getting a normal soldering iron to the other pads of the chip and producing a neat or reliable job on them. Maybe I'm just looking for any excuse to buy a new soldering station??
I agree with you on the recommended loads there. For an isolated micro controller and LCD display (all isolated from the audio electronics) I don't think we should worry too much about noise performance
An LM317 would do everything needed.Chris
Last edited:
bigpandahk, mate I think you mistake my comment which was meant for someone else... sure this is a little familiar territory, but hey I admire your enthusiasm and patience (hmm interesting choice of words), Ian offered you the board freely, but yeah you gotta look both ways before crossing the road man, its what brings you down every time.
no attaching a sink to the top of the chip is extremely unlikely to help, youde be lucky to dissipate 10% out the top, due to the thermal impedance of the package. the powerpad is there for a reason, on that thing you used it has nothing to do with heatsinking, some use it for a voltage reference. you are after DGN20 POWERPAD, powerpad is part of the package name, not just a trademark, without that the thing will come close to frying itself even without a load, hell, maybe even quicker without a load...
you cant continue to play this loose mate, especially given in a few weeks you are going to be putting 100VDC across the output of your BIII with only a couple zeners and a hopefully short free board between them. that sort of currents and voltage can sizzle and pop components. you need to stop adding new stuff...now! so that the system is stable, known and all fully functional. otherwise you could easily see the dac and several hundred worth of resistors go up in smoke before you know it and to replace them wont be as cheap...
glad its got a hole Ian, that works quite well. the same technique I/we used for a DGN14 (LT3032) and lt1963A/1764A, but well its a lot smaller and has 2 holes in islands separated by 0.5mm one ground, one neg rail... the Ti should be a walk in the park. I hadnt spotted the holes, but my comments stand...for other reasons
no attaching a sink to the top of the chip is extremely unlikely to help, youde be lucky to dissipate 10% out the top, due to the thermal impedance of the package. the powerpad is there for a reason, on that thing you used it has nothing to do with heatsinking, some use it for a voltage reference. you are after DGN20 POWERPAD, powerpad is part of the package name, not just a trademark, without that the thing will come close to frying itself even without a load, hell, maybe even quicker without a load...
you cant continue to play this loose mate, especially given in a few weeks you are going to be putting 100VDC across the output of your BIII with only a couple zeners and a hopefully short free board between them. that sort of currents and voltage can sizzle and pop components. you need to stop adding new stuff...now! so that the system is stable, known and all fully functional. otherwise you could easily see the dac and several hundred worth of resistors go up in smoke before you know it and to replace them wont be as cheap...
glad its got a hole Ian, that works quite well. the same technique I/we used for a DGN14 (LT3032) and lt1963A/1764A, but well its a lot smaller and has 2 holes in islands separated by 0.5mm one ground, one neg rail... the Ti should be a walk in the park. I hadnt spotted the holes, but my comments stand...for other reasons
Last edited:
I find this to do miracles sometimes. It is also right to say that, if not used properly could be a big nightmare. In this case where soldering needed to be done on the bottom flat surface of the chip it might do a really good job while trying to hand solder. Obviously if used with stencil and in reflow oven than that is what is meant for, but for hand work, one has to be careful. And it does not need to be this specific brand
Solder Paste, SMD & SMT Rework Solder Paste, Long Shelf Life, Ships Fresh!
Solder Paste, SMD & SMT Rework Solder Paste, Long Shelf Life, Ships Fresh!
I thought I had seen that hole when looking at the prototype pcbs when you posted photos. I'm still not sure about getting a normal soldering iron to the other pads of the chip and producing a neat or reliable job on them. Maybe I'm just looking for any excuse to buy a new soldering station??
I agree with you on the recommended loads there. For an isolated micro controller and LCD display (all isolated from the audio electronics) I don't think we should worry too much about noise performance
Chris
Yes, it needs a bit skill on soldering the QFN. I think the SMT rework station is a necessary though the thermal pad no problem for a normal iron.
This youtube may help:
How to Solder QFN MLF chips Using Hot Air without Solder Paste and Stencils - YouTube
Ian
I agree.
bigpandahk, The calcs I did above were to show how easy it would have been to calculate the temps expected before just trying it and waiting to see what would happen.
Also, what were you planning to use as an input to these regs?
Hi Ian,
I didn't mean to ask for anything free and never did in this DIY community, I am more than happy to pay for anyone's desgin work. Since this regulator pcb is not on the curent GB that's why I ask to see if it would be available, I like the size of your regulator PCB. I knew there is a EVM for only 20$ but it is bigger, 2.5" x 1.6". Anyhow, I will order a EVM kit to play with. No hard feelings and I enjoy your work very much.
Cheers
Last edited:
quap : "its what brings you down every time." - it's really true, I failed a lot and lost lot of money but still not learnt enough from experience
I will reconsider the whole power scheme for the project :
SSLV(1) for Buffalo, OTTO II
SSLV(2) for WaveIO, SideCar
7805 or LM317 for relays and battery management board
7812 for Arduino with LCD
TPS7A4700(1) for DSD - USB converter (with Ian's PCB)
TPS7A4700(2) for Teleporter
Batteries for i570, FIFO and SPDIF
The size of the casing will limit the quantity of transformer to be used. I have to combine them, may be one for the SSLV(s), one for the 78xx and the third one for the TPS. Including the two for NTD1, I hope there is enough space for 5 transformers.
I will reconsider the whole power scheme for the project :
SSLV(1) for Buffalo, OTTO II
SSLV(2) for WaveIO, SideCar
7805 or LM317 for relays and battery management board
7812 for Arduino with LCD
TPS7A4700(1) for DSD - USB converter (with Ian's PCB)
TPS7A4700(2) for Teleporter
Batteries for i570, FIFO and SPDIF
The size of the casing will limit the quantity of transformer to be used. I have to combine them, may be one for the SSLV(s), one for the 78xx and the third one for the TPS. Including the two for NTD1, I hope there is enough space for 5 transformers.
quap : "its what brings you down every time." - it's really true, I failed a lot and lost lot of money but still not learnt enough from experience
I will reconsider the whole power scheme for the project :
SSLV(1) for Buffalo, OTTO II
SSLV(2) for WaveIO, SideCar
7805 or LM317 for relays and battery management board
7812 for Arduino with LCD
TPS7A4700(1) for DSD - USB converter (with Ian's PCB)
TPS7A4700(2) for Teleporter
Batteries for i570, FIFO and SPDIF
The size of the casing will limit the quantity of transformer to be used. I have to combine them, may be one for the SSLV(s), one for the 78xx and the third one for the TPS. Including the two for NTD1, I hope there is enough space for 5 transformers.
bigpandahk,
Try this way to soldering the QFN, I believe you will make it
.How to Solder QFN MLF chips Using Hot Air without Solder Paste and Stencils - YouTube
Ian
bigpandahk,
Try this way to soldering the QFN, I believe you will make it
How to Solder QFN MLF chips Using Hot Air without Solder Paste and Stencils - YouTube
Ian
Thanks Ian.
hey man, no need for the bloody head, youre smart enough and from what ive seen skilled enough with the solder station, just not patient enough. I dont know if you do a lot of this stuff after long days at work/college etc? but many of the mistakes ive seen are mostly from rushing, or just thinking things will be OK, without KNOWING they will be ok. none of this is rocket-science, you can do it, but as above, you need to break it down into sub circuits, plan out what the voltage drop will be, trace out the ground connections, make each one of those sub-circuits work in isolation, then load tested with a suitable resistor to simulate what the regulator, or amplifier will see in use.
then and only then connect it to the real load, leave it connected, try and predict what the voltage drop should be based on the predicted load current, see how close that is to reality. THEN plan your next move, not immediately, chill on it, listen to some tunes and move on.
the above is a very good start, draw out the connections and what the expected input voltage, output voltage and difference between those is to give you the heat (not here, a simple flow chart type program will be) then you can use that to work out hoew much heat will need to be burned off, or if heat is not a concern, what the power/current budget is for that leg of the circuit etc etc
now i'm buggered,I gotta crash, you can do this, but youve gotta be smart about it, its not mechano
then and only then connect it to the real load, leave it connected, try and predict what the voltage drop should be based on the predicted load current, see how close that is to reality. THEN plan your next move, not immediately, chill on it, listen to some tunes and move on.
the above is a very good start, draw out the connections and what the expected input voltage, output voltage and difference between those is to give you the heat (not here, a simple flow chart type program will be) then you can use that to work out hoew much heat will need to be burned off, or if heat is not a concern, what the power/current budget is for that leg of the circuit etc etc
now i'm buggered,I gotta crash, you can do this, but youve gotta be smart about it, its not mechano
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.