My preferred type of enclosure is an open-top aluminum extrusion. Here's an example from one of my suppliers
If anyone cares, I can design a 3D printable enclosure and publish the files when the board reaches my desk
I can probably cobble together enclosure design for laser cutters too (I have a glowforge and have made some box designs). I might even be convinced to make some for folks here as long as it's not a crazy number.
**slaps roof of forum** - this bad boy can fit so many enclosure solutions...
I have a laser cutter here too and and probably help with the load too.
I will say though, that supplier that Phil posted above looks really good for something else I'm working on.
The two unlabeled 3 pin female connectors (white) in the upper left corner of each regulator section above each large inductor are used for...? Sorry if this has already been covered.
The two unlabeled 3 pin female connectors (white) in the upper left corner of each regulator section above each large inductor are used for...? Sorry if this has already been covered.
It hasn't been covered. Those sockets are for inserting resistors which can pull the regulated voltage up or down. A resistor in the top two holes will lift the voltage up, while a resistor in the bottom two holes will pull the voltage down.
I see no reason the beta chips can't have a COG do the TMDS encoding on the fly, then another cog do the actual streaming. Right now it's streaming 1 bit per clock, which means you get 10 clocks to calculate the TMDS data for an RGB tuplet.
Chip,
I have both the Professional Development Boards (Propeller and BasicStamp) and I have found the 8 under-board rubber pads enough, I think there is no need for enclosures.
But I'll really like if you can add a few (let say four) MikroElektronica style click connectors (https://mikroe.com/click).
I've noted that beside Microchip (which can be reasonable because they are partners) also other dev boards (fpga, arduino compatible, ...) start using this standard because of the numerous click-boards available.
Perhaps your hdmi, vga, usb, ... daughter-boards can also use this standard. Both Parallax and MikroE can have benefits from this choice.
PS. Put me down for one of these boards. Pls take in account Italian/European custom fees.
MPS also do a 3A version (Mpm3632c) which isn't too much more expensive but adds a fair bit of board area.
Nice looking parts.
Not quite sure what your board area comment refers to ? - Digikey shows both MPM3620A & Mpm3632c in 20-QFN (5x3)
I've avoided too much research into integrated inductors, as the price premium is high, but they do have a low BOM, and a small PCB area, so maybe I should include some in the list.
P2 customers may not be as price-paranoid as some I'm used to...
MPS also do a 3A version (Mpm3632c) which isn't too much more expensive but adds a fair bit of board area.
Reading their data more carefully, shows the MPM3620A only has a low rated EN pin, so needs a zener clamp if fed > 6.5V, - fixed in the Mpm3632c
MPM3620A / Mpm3632c are similar package, but not pin compatible.
I also see TI have some that come in a family range, pin compatible, with a 0.65mm lead spacing.
TI Low Voltage DC-DC modules
TPS82084 - Integrated Inductor 2A 2.5~6V 0.8~6V LGA 3.0x2.8x1.5mm SS:Fixed EMI Tested, Enable, Light Load Efficiency, Output Discharge, Power Good -40 to 125 $1.44 | 1ku 0.017mA
TPS82085 - Integrated Inductor 3A 2.5~6V 0.8~6V LGA 3.0x2.8x1.5mm SS:Fixed EMI Tested, Enable, Light Load Efficiency, Output Discharge, Power Good -40 to 125 $1.65 | 1ku 0.017mA
Between families : Similar package, but LV is not pin compatible with MV package
TI medium voltage DC-DC modules (all pin compatible)
TPS82150 - Integrated Inductor 1A 3~17V 0.9~6 LGA 3.0x2.8x1.5mm SS:Adjustable EMI Tested, Enable, Light Load Efficiency, Output Discharge, Power Good, Tracking -40 to 125 $1.50 | 1ku 0.02mA
TPS82140 - Integrated Inductor 2A 3~17V 0.9~6 LGA 3.0x2.8x1.5mm SS:Adjustable EMI Tested, Enable, Light Load Efficiency, Output Discharge, Power Good, Tracking -40 to 125 $1.65 | 1ku 0.02mA
TPS82130 - Integrated Inductor 3A 3~17V 0.9~6 LGA 3.0x2.8x1.5mm SS:Adjustable EMI Tested, Enable, Light Load Efficiency, Output Discharge, Power Good, Tracking -40 to 125 $1.82 | 1ku 0.02mA
Given Chip's requirements of "We would like 3 amps.", that TPS82130 looks attractive, gives higher current than Torex, and an easier footprint & good PCB design. (tho I think reflow only)
I see TI also not have a store, and you can buy any of these, strangely, the 3A one is cheaper than the 2A one in their store. TPS82130SILR 3-A 17V TI Store Price: 750 - 999 $ 1.86 1000 - 9999 $ 1.82
TPS82140SILT 17V Input 2A 750 - 999 $ 1.99 1000 - 9999 $ 1.90
TPS82150SILT 17V Input 1A 750 - 999 $ 1.81 1000 - 9999 $ 1.73
Looking further afield, uncovers a MUN12AD03-SH - claims 4.5~16V, 3A, and package quite similar to TI's (only with 0.8mm pad pitch, 3.5 x 3.5mm)
That seems to have a low price, but somewhat cursory specs.
I have two UM232H modules from FTDI (it's their FT232H development board).
The FT232, for it's highest speed mode called '245 synchronous FIFO' (able to transfer 40 MBYTES/s) requires the CPU to be synchronized to its own 60MHz clock output.
That is one of the reasons I need the board to allow an external clock. One example of such high transfer rate could be to capture the whole 4GB XORO output in 409.6 seconds (6 minutes and 49.6 seconds) instead of waiting around 5 hours if we use a simple 1MB/s FT245/FT232R usb cable.
Also there are other weird frequencies I want to use (and they are not exact multiple of the 20MHz crystal that I guess will be installed on board). Using the VCO/PLL integer divider is not an option as frequency needs to be exact ...
BTW, I have too a 10 MHz atomic clock (rubidium) and I would be interested in playing with P2 as frequency/pulse counter/generator, ...
Though I doesn't had enough time to study the datasheet provided by FTDI at full, in case of FT245 Synchronous FIFO Interface Mode as depicted at page 26 and 27, IMHO, you'll have a lot of good chances to use it without being forced to derivate P2's 20 MHz clock input (or any other suitable frequency), from FT232H's 60 MHz clock output.
To the extent of what I've read so far, you could run the P2 at 120 MHz or, even better, at 240 MHz (because there is a good chance for a HDMI sink to recognize meaningfull TMDS data @ 24 MHz). Doing so, it'll be easy to command one of the smart pins, to generate a solid 12 MHz, 3.3 V clock output, to feed FT232H clock input.
P2 @ 240 MHz would sport 20 x 4.1667 nS (granularity) timing slots, for you to sync Cogs and Streamers, and construct almost any tightly-timed output signaling or data, or recover timing-synched data.
Again, can be a bit of chalenging to sync them, thus its advisable to have good instruments at reach (oscilloscope/logic analyzer), but, at the end of the day, the results can be rewarding too.
Next to the VIO headers are local, quiet 3.3V LDO regulators. They will be useful for quiet analog on sets of 8 pins, each.
Those jumpers can also be opened up for current measurements, from either the main VIO supply, which is adjustable, or the local 3.3V regulator.
The main 1.8V and 3.3V switching regulators are both adjustable, up and down. the top two jumpers can also be removed for current measurements on either main supply.
How quiet are those regulators ? What is the part number ?
It may be useful to allow a second footprint alongside, for a truly quiet voltage reference regulator. Not all pins need this, maybe fit 1 or 2 ?
Digikey finds 182 stocked (3v, 3v3), when < 10ppm/°C ticked, many are in SOT23-5 or -6 pins
Next to the VIO headers are local, quiet 3.3V LDO regulators. They will be useful for quiet analog on sets of 8 pins, each.
Those jumpers can also be opened up for current measurements, from either the main VIO supply, which is adjustable, or the local 3.3V regulator.
The main 1.8V and 3.3V switching regulators are both adjustable, up and down. the top two jumpers can also be removed for current measurements on either main supply.
How quiet are those regulators ? What is the part number ?
It may be useful to allow a second footprint alongside, for a truly quiet voltage reference regulator. Not all pins need this, maybe fit 1 or 2 ?
Digikey finds 182 stocked (3v, 3v3), when < 10ppm/°C ticked, many are in SOT23-5 or -6 pins
An appeal of the 3.000V ref, it is can feed from the 3.3V regulator nearby, for lowest noise Vin.
10mA output should be enough to drive DACs for self testing ?
Those 3.3V LDOs are ON Semi part# NCP114AMX330TCG.
From the datasheet:
•Low Dropout: 135 mV Typical at 300 mA
•±1% Accuracy at Room Temperature
•High Power Supply Ripple Rejection: 75 dB at 1 kHz
•Thermal Shutdown and Current Limit Protections
•Stable with a 1 uF Ceramic Output Capacitor
We use TPS7A49xx series. I went for a bit of a hunt, LP5907 looks reasonable, noise is an order of magnitude down vs NCP1114
Input Voltage Range: 2.2 V to 5.5 V
• Output Voltage Range: 1.2 V to 4.5 V
• Stable With 1-µF Ceramic Input and Output
Capacitors
• No Noise Bypass Capacitor Required
• Remote Output Capacitor Placement
• Thermal-Overload and Short-Circuit Protection
• –40°C to 125°C Operating Junction Temperature
• Low Output Voltage Noise: < 6.5 µVRMS
• PSRR: 82 dB at 1 kHz
• Output Voltage Tolerance: ±2%
• Very Low IQ (Enabled): 12 µA
• Low Dropout: 120 mV (typical)
We use TPS7A49xx series. I went for a bit of a hunt, LP5907 looks reasonable, noise is an order of magnitude down vs NCP1114
...
That LP5907 may be a better default 3v3 regulator, but I'd still suggest using a 3.000V reference in a couple of places, as the temperature drift of the LP5907 is not as good as the MAX6071BAUT30, and if Parallax want to temperature qualify their DACs they will need a good supply.
I intend to use LDOs after the switching regulator to improve the load response and noise. To maintain full 3.3V this requires the switching reg output up to 200mV or so higher, the less the better though. But I'm making that optional on my new boards so that we can go straight from the switching reg which is set for the correct voltage, or via the LDO, or no switching reg and just the LDO for the 3.3V supply. The 1.8V will always need a switcher due to the current and voltage difference. options options options. If the 1.8V supply has a 1.9V switcher and if it was supplying 1A then that is still only 100mW power dissipation in the 1.8V LDO which of course needs to work at these levels. I have always done this in the past with 5V switchers to 3.3V LDOs.
Comments
**slaps roof of forum** - this bad boy can fit so many enclosure solutions...
I have a laser cutter here too and and probably help with the load too.
I will say though, that supplier that Phil posted above looks really good for something else I'm working on.
It hasn't been covered. Those sockets are for inserting resistors which can pull the regulated voltage up or down. A resistor in the top two holes will lift the voltage up, while a resistor in the bottom two holes will pull the voltage down.
Put a lot of COGS to work, and making a dynamic display looks possible, though likely not practical.
Can't wait! : ]
I plan to explore HDMI, but will likely run component most of the time.
I have both the Professional Development Boards (Propeller and BasicStamp) and I have found the 8 under-board rubber pads enough, I think there is no need for enclosures.
But I'll really like if you can add a few (let say four) MikroElektronica style click connectors (https://mikroe.com/click).
I've noted that beside Microchip (which can be reasonable because they are partners) also other dev boards (fpga, arduino compatible, ...) start using this standard because of the numerous click-boards available.
Perhaps your hdmi, vga, usb, ... daughter-boards can also use this standard. Both Parallax and MikroE can have benefits from this choice.
PS. Put me down for one of these boards. Pls take in account Italian/European custom fees.
As others have asked, is there a signup sheet?
What is your Voltage-in target range, and do you need PGood ?
The input is 5 volts from the USB. We do not need a power-good indication. We would like 3 amps.[/quote]
Did you select a better switcher yet ?
Seeing threads like https://forums.parallax.com/discussion/169191/nybble-the-cat#latest reminds me why I prefer wider range switchers...
I'll add another one to the list, higher voltage and smaller package than SO8, eliminates external compensation parts
..from Diodes inc,
AP65355FN-7 4.5V~18V Vo 0.76V to 6V 3A PGood 650kHz 1.5% U-DFN3030-10 (3x3) $0.487/3k (Mouser $0.426/3k) Data May 2017
Needs BS, has SS, 90mΩ 57mΩ FETS
UVLO Threshold VIN Rising 3.6 3.85 4.1 V
and the older list was :
NCP1593AMNTWG OnSemi 3A SYNC 10-DFN 3,000 $0.35/3k 4V~5.5V 0.6V 3A 1MHz -40°C ~ 85°C 10-DFN (3x3)
FETs 90 mΩ/60 mΩ Hiccup, Soft−Start, PGood, internal compensation, no Boost C needed, 100% Duty support.
Older tech, but wider Vin
NCP3170A/B OnSemi 4.5~18V, 3A, PGood, 33c/2.5k
FETs 100 mΩ / 29 mΩ typs at 4.5V SO8 Eval is NCP3170AGEVB
https://monolithicpower.com/pub/media/document/MPM3620A_r1.0.pdf
MPS also do a 3A version (Mpm3632c) which isn't too much more expensive but adds a fair bit of board area.
Nice looking parts.
Not quite sure what your board area comment refers to ? - Digikey shows both MPM3620A & Mpm3632c in 20-QFN (5x3)
I've avoided too much research into integrated inductors, as the price premium is high, but they do have a low BOM, and a small PCB area, so maybe I should include some in the list.
P2 customers may not be as price-paranoid as some I'm used to...
So it would be nice if that was easy to do.
Either way, I WANT ONE !!!!
Bean
Reading their data more carefully, shows the MPM3620A only has a low rated EN pin, so needs a zener clamp if fed > 6.5V,
MPM3620A / Mpm3632c are similar package, but not pin compatible.
I also see TI have some that come in a family range, pin compatible, with a 0.65mm lead spacing.
Given Chip's requirements of "We would like 3 amps.", that TPS82130 looks attractive, gives higher current than Torex, and an easier footprint & good PCB design. (tho I think reflow only)
I see TI also not have a store, and you can buy any of these, strangely, the 3A one is cheaper than the 2A one in their store.
TPS82130SILR 3-A 17V TI Store Price: 750 - 999 $ 1.86 1000 - 9999 $ 1.82
TPS82140SILT 17V Input 2A 750 - 999 $ 1.99 1000 - 9999 $ 1.90
TPS82150SILT 17V Input 1A 750 - 999 $ 1.81 1000 - 9999 $ 1.73
TPS82084SILR 2-A, High-Efficiency Step-Down Converter Module with Integrated Inductor 750 - 999 $ 1.51 1000 - 9999 $ 1.44
TPS82085SILR 3A, High-Efficiency Step-Down Converter Module with Integrated Inductor (low voltage) 750 - 999 $ 1.68 1000 - 9999 $ 1.65
High volume P2 users, who are certain they never need > 5V in, could change PCB layouts to the TPS82084/5, to prune a few cents.
There are eval boards for the TPS82130/40/50.
It looks like ferrite beads on IN and OUT sides of these, may be needed for conducted emissions pass.
That seems to have a low price, but somewhat cursory specs.
Has anyone ever used Cyntec modules ?
The FT232, for it's highest speed mode called '245 synchronous FIFO' (able to transfer 40 MBYTES/s) requires the CPU to be synchronized to its own 60MHz clock output.
That is one of the reasons I need the board to allow an external clock. One example of such high transfer rate could be to capture the whole 4GB XORO output in 409.6 seconds (6 minutes and 49.6 seconds) instead of waiting around 5 hours if we use a simple 1MB/s FT245/FT232R usb cable.
Also there are other weird frequencies I want to use (and they are not exact multiple of the 20MHz crystal that I guess will be installed on board). Using the VCO/PLL integer divider is not an option as frequency needs to be exact ...
BTW, I have too a 10 MHz atomic clock (rubidium) and I would be interested in playing with P2 as frequency/pulse counter/generator, ...
Though I doesn't had enough time to study the datasheet provided by FTDI at full, in case of FT245 Synchronous FIFO Interface Mode as depicted at page 26 and 27, IMHO, you'll have a lot of good chances to use it without being forced to derivate P2's 20 MHz clock input (or any other suitable frequency), from FT232H's 60 MHz clock output.
To the extent of what I've read so far, you could run the P2 at 120 MHz or, even better, at 240 MHz (because there is a good chance for a HDMI sink to recognize meaningfull TMDS data @ 24 MHz). Doing so, it'll be easy to command one of the smart pins, to generate a solid 12 MHz, 3.3 V clock output, to feed FT232H clock input.
P2 @ 240 MHz would sport 20 x 4.1667 nS (granularity) timing slots, for you to sync Cogs and Streamers, and construct almost any tightly-timed output signaling or data, or recover timing-synched data.
Again, can be a bit of chalenging to sync them, thus its advisable to have good instruments at reach (oscilloscope/logic analyzer), but, at the end of the day, the results can be rewarding too.
Henrique
How quiet are those regulators ? What is the part number ?
It may be useful to allow a second footprint alongside, for a truly quiet voltage reference regulator. Not all pins need this, maybe fit 1 or 2 ?
Digikey finds 182 stocked (3v, 3v3), when < 10ppm/°C ticked, many are in SOT23-5 or -6 pins
A good candidate looks like
MAX6071BAUT30+T Maxim 3V SOT23-6 583 stk $1.30560.3k Io=10mA ±0.08% 8ppm/°C 4.6µVp-p 7.8µVrms Vi 3.2 V ~ 5.5 V 300µA -40°C ~ 125°C
MAX6070AAUT30+T Maxim 3V SOT23-6 1,029 stk $2.02368/3k Io=10mA ±0.04%
An appeal of the 3.000V ref, it is can feed from the 3.3V regulator nearby, for lowest noise Vin.
10mA output should be enough to drive DACs for self testing ?
Those 3.3V LDOs are ON Semi part# NCP114AMX330TCG.
From the datasheet:
•Low Dropout: 135 mV Typical at 300 mA
•±1% Accuracy at Room Temperature
•High Power Supply Ripple Rejection: 75 dB at 1 kHz
•Thermal Shutdown and Current Limit Protections
•Stable with a 1 uF Ceramic Output Capacitor
Input Voltage Range: 2.2 V to 5.5 V
• Output Voltage Range: 1.2 V to 4.5 V
• Stable With 1-µF Ceramic Input and Output
Capacitors
• No Noise Bypass Capacitor Required
• Remote Output Capacitor Placement
• Thermal-Overload and Short-Circuit Protection
• –40°C to 125°C Operating Junction Temperature
• Low Output Voltage Noise: < 6.5 µVRMS
• PSRR: 82 dB at 1 kHz
• Output Voltage Tolerance: ±2%
• Very Low IQ (Enabled): 12 µA
• Low Dropout: 120 mV (typical)
That LP5907 may be a better default 3v3 regulator, but I'd still suggest using a 3.000V reference in a couple of places, as the temperature drift of the LP5907 is not as good as the MAX6071BAUT30, and if Parallax want to temperature qualify their DACs they will need a good supply.