Lattice iCE40HX1K FPGA

[`FRANK.C]

Over the holidays I have been looking around to see what FPGA's are in stock.
One sticks out, the Lattice CE40HX1K-TQ144 which is in current production. It is in stock from a number of suppliers and a couple of suppliers have a few thousand in stock.
I suspect that it is not as sought after as outers because its resources are pretty meager.

1280 Logic cells
64k bit BRAM
1 PLL
No multipliers

Has it enough resources to produce a no frills standards converter?
I think there is enough memory to do 2 line stores and the time redistribution memory.
We can work with 1 PLL

But would there be enough Logic cells considering that they will have to be used to do the multiplier(s)
Does a 8 X 8 multiplier use many cells?

The programming of the TVP5150 and modulators could be farmed out to a microcontroller to save cells.

I have tried out its programming tools iCEcube2. They are basic but workable.

Frank

[ppppenguin]

I implemented a standards converter in a Xilinx XC2S200 which has just over 5000 logic cells.  I think I was being fairly clever about the interpolation multiplier, multiplexing all 4 lines through it at 4x clock. The design used SDRAM framestores. All interpolation was after the framestores since I had enough memory bandwidth to make it look like a multiport RAM. If you're using block RAM to do  internal linestores then you're likely to run out of that resource before anything else.

The mapping report said:
Logic Distribution:
    Number of occupied Slices:                      1,072 out of  2,352  45%
    Number of Slices containing only related logic:  1,072 out of  1,072  100%
    Number of Slices containing unrelated logic:        0 out of  1,072    0%
        *See NOTES below for an explanation of the effects of unrelated logic
Total Number of 4 input LUTs:        1,198 out of  4,704  25%
      Number used as logic:                    1,041
      Number used as a route-thru:                145
      Number used as Shift registers:              12
  Number of bonded IOBs:          202 out of    284  71%
      IOB Flip Flops:                            149
      IOB Latches:                                  8
  Number of Block RAMs:              8 out of    14  57%
  Number of GCLKs:                  3 out of      4  75%
  Number of GCLKIOBs:                3 out of      4  75%
  Number of DLLs:                    2 out of      4  50%

I think a synthesised 8x8 multiplier uses about 64 cells. Needs a fair bit of internal pipelining to work at video speeds.

I've attached all the VHD files from my design but I think I sent you these long ago. It will take a fair but of dredging my own memory banks to remember much more.

This was the design I later adapted to do 405 to 625 conversion. The hardware is a board I designed for a professional cliient some years ago. I happen to have a few around.

PS: Don't try to design a multiplier yourself. I've done this on older FPGAs before there were tools like Coregen to do the job for you. It's a lot of work to design and debug.

PPS: I've totally lost touch with the FPGA market. Haven't done any real design work for years. Xilinx and Altera have both been bought by huge companies (Intel and AMD) who I suspect aren't interested in small devices anymore. So you're loking at 2nd tier makers like Lattice.

[`FRANK.C]

Hi Jeffrey
Thanks for replying.
Looks very tight then.

Frank

[ppppenguin]

Roundup of low cost FPGA dev kits: https://www.joelw.id.au/FPGA/CheapFPGADevelopmentBoards

Try not to use obsolete parts such as Xilinx Spartan 3 or even Spartan 6 which is no longer current. The Arty Spartan 7 board is relatively cheap. https://digilent.com/shop/arty-s7-sparta...ent-board/

[`FRANK.C]

Hi Jeffrey
That a good list of development boards. Trouble is many are out of stock or preorder.
The ones in stock tend to be at the smaller end of the scale.
Even the boards from China are scarce and have increased a lot in price.

Frank

[ppppenguin]

Might be worth looking for Arduino and Raspberry Pi FPGA add-on boards. The PIF2 here might work: https://bugblat.com/products/pif/index.html

You don't have to use them with their Arduino or R-Pi hosts but it may be convenient to do so.

[`FRANK.C]

I have used the iCEcube2 tools to synthesize a standards converter in the CE40HX1K
I altered the Hedghog files that I had considerably.  I pruned them back to the essential items.
I ended up with a 2 line interpolater, 2 multipliers, a 1 line FIFO, most of the rest of the memory is used in the time redistribution memory.
It wont do the programming of the TVP5010 or modulators.
The video output is not over sampled.

Resource Usage Report for HEDGHOG_L_1

Mapping to part: ice40hx1ktq144
Cell usage:
GND               4 uses
SB_CARRY           342 uses
SB_DFF             52 uses
SB_DFFE           203 uses
SB_DFFESR         23 uses
SB_DFFESS         5 uses
SB_DFFR           2 uses
SB_DFFSR           88 uses
SB_DFFSS           1 use
SB_GB             6 uses
SB_RAM1024x4       2 uses
SB_RAM2048x2       12 uses
VCC               4 uses
pll               1 use
SB_LUT4           749 uses

I/O ports:       18
I/O primitives:       18
SB_GB_IO         1 use
SB_IO             17 uses

I/O Register bits:                     0
Register bits not including I/Os:     374 (29%)

RAM/ROM usage summary
Block Rams : 14 of 16 (87%)

Total load per clock:
  HEDGHOG_L_1|CK_27_MHZ: 1
  HEDGHOG_L_1|CK_PLL_216_MHZ_inferred_clock: 1
  in_625|CLK_derived_clock: 134

@S |Mapping Summary:
Total  LUTs: 749 (58%)

Distribution of All Consumed LUTs = LUT4
Distribution of All Consumed Luts 749 = 749

Mapper successful!

I have looked at it using the RTL viewer and it appears to be all there.
It looks promising but I take all of the above with a pinch of salt as I know that just missing out on one signal can result in a huge chunk of the circuit not getting synthesized.

Until I try it in a circuit I wont really know if it will work
So I have ordered a development board they are not expensive.
https://www.olimex.com/Products/FPGA/iCE...e-hardware

Frank

[Jac Janssen]

Interesting work Frank.

Not many inexpensive (larger) FPGAs seem to be available.

Jac

[`FRANK.C]

Hi Jac
Yes FPGA's have gone very scarce.
Even the bigger brother of this one the ICE40HX4K-TQ144 is in short supply.

I just noticed that I left out the 'I' at the beginning of the part number in the title of this thread.

Frank

[`FRANK.C]

Thanks for correcting the title 

Frank